Abstract

This study investigates the effect of competition, size, and development indicators on efficiency using the panel data of 1137 BRICS (Brazil, Russia, India, China, South Africa) banks over sixteen years 2000-2015. The core finding of the two-step system generalized methods of moments (2GMM) are: (i) Both competition and size have a significant positive association with the efficiency of BRICS banks. (ii) No significant association between inflation and efficiency is observed; whereas, economic progression (gross domestic products) has a positive effect on the efficiency of revenue and inverse effect on cost efficiency. (iii) The interim term size and competition have a homogeneous effect on each type of efficiency; however, the outcomes are heterogeneous when efficiency measures change. Finally, the study evidences the nonlinear relationship of competition, size, and development indicators of BRICS banks.

1. Introduction

The impact of competition on banks' performance is a long-standing debated issue among researchers ¹ in the last few years. The economic giants say Brazil, Russia, India, China, South Africa (henceforth BRICS) banks become worth considering as a field of study due to their successful survival in the global financial crisis 2007-2008 and fewer banking failure ^{2, 3}. The competition was not treated as a critical factor for the financial sector as most of the nations were not interested in following the standard competition policy. However, in the last few decades, the financial market crisis forced the regulators and policymakers to strengthen the competition policy at the national level ⁴. Mentioning competition as an essential factor of efficiency of banks, ⁵ summarizes six reasons why competition is worth considering factor for the financial system from their literature survey: firstly, to access the financial services for firms and households; secondly, for proper activity performance of financial sector; thirdly, for stability; fourthly, to confirm efficiency; fifthly, to settle a stable market rate and improved monetary policy; and finally, for economic growth and industrial development. Thus, the literature points out the importance of competition in banks' efficiency and the country's economic progress.

BRICS, the fastest growing economy (see Figure 1, the GDP over the years) in the world, contributes almost one-fourth of the world GDP and consists of about 42% of the world population ⁶. Economic progress with a large segment of the world population draws the researchers attention to unfold their hidden momentum of fast growth and fewer failure in the crisis ². Rapid growth of BRICS nations, fuels researchers to investigate the relationship between risk, competition, and performance ^{2, 7} and concentration risk and performance. However, rare work observed the connection between the competition and efficiency; and competition and development indicators of this region ³. Figure 1 and Table 1 illustrate the snapshot of GDP and the market concentration of BRICS nations.

This paper aims to revisit the relationship between competition and efficiency of banks and extend the work of ⁴ on the developed and emerging financial markets. Most of the results examining the competition and performance are based on single countries ^{1, 8} and cross countries ^{4, 9}. Study of BRICS as a region observed in very few research works ^{3, 5}. This literature gap builds our interest to investigate the effect of competition and economic progress on the BRICS nation. This study grounded to answer immediate questions such as (i) How competition and concentration affect banks' efficiency? (ii) How are efficiency and development indicators related? (iii) Do market competition and development indicators heterogeneously affect the efficiency of banks? This paper aims to contribute in the following manner: First, to shed light on the relationship between efficiency and competition (See in Table 1 different inverse proxy measure competition of BRICS banks) and efficiency and development indicators of BRICS banks. Second, to examine the linear and quadratic effect of competition and development indicators that rarely addressed in the literature ⁶. Finally, to extend the work by examining the nonlinear impact of development indicators and market concentration, and by showing the size effect of explanatory variables concentration and development indicators to extend the model of ^{1, 4, 10}.

The rest parts of the study are organized as follows. Section 2 describes the relevant literature of the study; Sections 3 illustrates the data & variables description. Section 4 describes the study's empirical methodology. Section 5 presents the analysis results explaining the relationship between competition and development indicators with nonlinear and quadratic effects. Finally, Section 6 contains the concluding remarks of the study.

2. Literature Review

Two dominant hypotheses explain the interrelationship between risk and efficiency. These are the 'competition efficiency' hypothesis and 'competition-inefficiency' hypothesis. Following the efficient structure of ¹¹, the competition-efficiency hypothesis illustrated by ¹². The authors pinpoint that competition acts as an external force that stimulates banks to serve at a low cost and become cost-efficient. However, size gives banks extra benefits to cope with competition, operates the business in a cost-efficient manner, and have more market power. In a nutshell, the competition-efficiency hypothesis explains that competition leads banks to efficiency. Non-competitive market leads managers to enjoy a 'quiet life'^{1} that makes the banks cost-inefficient as managers become unconscious about cost control.

The 'competition-inefficiency' hypothesis illustrates the inverse of the competition-efficiency hypothesis concept. It explains that market competition decreases the efficiency of banks ¹². ¹² argue that competition-inefficiency becomes ostensible for several reasons. In the most competitive market, customers enjoy more choices and tend to switch banks to explore more benefits. So banks fall into the position of shorter bank-customer relationship that makes banks more unstable. In a competitive environment, banks are unwilling to involve more resources to have information regarding customers that require to hold existing customers and attract new customers, as they are not in a concern of relationship-building activities in the competitive market environment. Thus in a competitive market, banks need to incur more cost to hold old customers and attract new customers due to information asymmetry and aggressive market effort. Empirical evidence also supports the adverse effect of market competition on the efficiency of banks. For example, ^{13, 14, 15}, among others, show the adverse effect of market competition over the efficiency of banks. ¹⁵ point out that with the increase of new entrants, the efficiency of banks decreases.

Delving the relationship between efficiency and competition, ¹⁶ opine that market competition inversely related to banks' efficiency, whereas the positive association observed a link between market concentration and efficiency. Opposing this view, ¹⁷ depicts the positive rapport between market competition and banks' efficiency. The authors opine that market competition exerts a significant positive influence on banks' cost and technical efficiency. Again ¹⁸ mention that there is no significant relationship between market competition and banks' efficiency. The authors comment that market competition cannot be detrimental to banks' efficiency gain from the Chinese banking industry survey. But evidence of a significant relationship between market competition and efficiency are also available. ¹⁹ show the meaningful relationship between market competition and efficiency of banks. The authors stress that with the increase of efficiency, banks can achieve more market power, and market power also plays a conducive role in enhancing banks' efficiency. Thus literature supports the evidence of the relationship between competition and efficiency of banks. In examining the size effect, ²⁰ comments that there is insufficient evidence regarding the relationship between banks' size and efficiency. The study of ⁴ also evidenced the association between competition and economic performance from the cross-countries empirical data. ²¹ opines that a competitive market shows the slow economic growth of a nation. On the contrary, ²² depict the positive association between development indicators like GDP and inflation with banks' performance. To delve, the relationship hypotheses of the study are:

H1: Market competition and development indicators have a significant impact on the cost-efficiency of banks.

H2: Market competition and development indicators have a significant impact on the revenue efficiency of banks.

H3: Impact of competition and development indicators have a heterogeneous effect on the efficiency of banks.

3. Data, Variables Description of the Study

This investigation uses bank-level, industry-level, and macroeconomic control variables. Bank-level controlled variables are collected from the bank-scope database; industry-level and macroeconomic data are collected from the world bank database. The monetary units are expressed in the US dollar and constant price. Finally, we have 9988 bank years' data from 1137 BRICS banks after excluding the dataset's missing data. The following are the details of the collected data.

In selecting data, we consider three years of consecutive data set for consistency. In this section, firstly, we discuss how the competition is measured and then provide the description of other variables.

To determine the competition, we use three anti-proxy measures of competition and direct measurements of market power. These measures are Lerner index, Herfindahl– Hirschman Index for assets and Herfindahl– Hirschman Index for gross loan

Lerner index is a widely used measure of competition in a different study ^{23, 24, 25, 26}. Theoretically, the index value lies between zero (0) and one (1). The higher the index's value, the greater the market power (Pricing power) and less competitive conditions ^{6, 19}. We also opted for the Lerner index as a proxy measure of market power and inverse measure of competition. The index is given as:

Where LI_jt refers the Lerner of j year t time. P_jt stands for the price of output of banks; MC_jt refers to the marginal cost. The index value we used is collected from the world banks database for the periods.

By following the study of ^{1, 6, 27}, among others, we opted for Herfindahl-Hirschman Index as an alternative inverse measure of competition based on assets and gross loans. The index's value is zero (0), which means homogenous firms are operating in the market and vice versa. That means the lower the value, the less the market power and more competitive conditions.

The market share refers to the particular firm proportion to its total industry value; by adding the square term of the ratio, the index is derived. Using the formula, we determine two market power measures, the HHIA-Herfindahl-Hirschman Index for Assets and HHIL-Herfindahl-Hirschman Index for gross loans for the study.

We opt for SFA to determine cost and revenue efficiency using the Bankscope database's information. SFA is also used by different empirical study say ^{3, 4, 7, 28, 29}, among others. Using SFA FRONTIER - Version 4.1c, we determine the efficiency value of cost and revenue; and uses values derived in further examination of the relationship. The details of the estimation are given in Appendix A. Table 2 explains the detailed description and sources of all variables.

4. Research Framework

This study opted for the System Generalized Method of Moments (GMM) approach to investigate the effect of competition and development in the cost and revenue efficiency of BRICS banks. Using the panel data, we run the regression equations and observe the endogeneity and heteroskedasticity problems. System GMM suggested by ⁴³ and ⁴⁴ is applied for our dynamic panel data to address the endogeneity and unobserved heteroskedasticity and autocorrelation problems of the model ^{2, 30, 45, 46}. The empirical model of the study is structured as follows:

(1)

Where the dependent variable presents cost and revenue efficiency. In the Equation, 'i' subscript refers to the cross-sectional dimension across banks, and subscript

j, l, k indicates macro-economic, industry-level, and bank-level control variables, respectively. 't' denotes the time dimension (i.e., t = 2000, 2001, 2002,…., 2015). One year lagged dependent variable represented by Y_i,t-1.

The X_i,j,t are the macroeconomic control variables: the gross domestic products (GDP) and inflation. The X_i,j,t represents the industry level control variables: banking sector development (BSD) and Competition measures (LI, HHIA, HHIGL) at t period. The X_i,k,t present the banks level control variables of bank ‘i’ at ‘t’ period. Bank-level control variables are capital to total assets (Cap to TA), return on assets (ROA), size for the efficiency of cost, revenue diversification (RD), risk (NPL), return on assets (ROA), and size for the efficiency of revenue measures.

We extend our baseline model to address the size effect and nonlinear effect on bank competition and development indicator.

Assuming the heterogeneous behaviour of bank in a competitive environment and size, the extended model is as follows:

(2)

Where the variable Com and refer 'Competition measure' and 'the squared term of competition' and S indicates 'large and small bank size' (Large bank derived by the subtracting average industry assets from the particular bank asset whereas Small bank derived by subtracting particular bank asset from the average industry assets).

Small and Large banks labelled as Model I and Model II respectively in extended results. Product of size and competition (nonlinear effect of competition) denote by, Inverse competition measure (LI, HHIA, HHIGL) refers to the market concentration, and the greater value of coefficient indicates the highly concentrated market and less competitive. denote the interactive term between size and gross domestic products (nonlinear effect). A higher positive (negative) value of this coefficient refers to the increasing (decreasing) efficiency of different size banks through economic progress.

5. Analysis of the Study

This section discusses the descriptive statistics (see Table 3) of variables and evaluates the multicollinearity through the correlation matrix (see Table 4). Subsequently, we explain the empirical finding to the study regarding the impact of competition and development indicators on bank efficiency (see Table 5 - Table 6). Nonlinear and quadratic effect of development indicators and competition by considering size discuss in Table 7 and Table 8. Small and large-sized banks are labelled as Model I and Model II, respectively, in Table 7 and Table 8. All Tables (Table 5 - Table 8) present the regression results of the two-step system GMM by applying different competition measures.

Summary statistics of the variables are presented in Table 3. The mean value of Efficiency of cost is 6.94, and the efficiency of revenue is about 0.32. The competition measure LI shows the mean value of 0.1630 refers to the fact that the BRICS banks are less concentrated and more competitive. The average value of HHIA and HHIGL is also less than 0.20, which means that banks are gaining less market power over the asset and gross loan in the BRICS region. That means the competitive condition exists in the bank. The average value of LI, HHIA, HHIGL of the latest three years is getting less than the overall mean value of concentration in all countries (see Appendix B). Thus it is confirmed that the market competition of BRICS countries is increasing day by day.

The average of small size and large size banks are 11593 and 11594 million USD, respectively. Capital to total assets that means a proportion of equity to total asset shows the average figure of about 22.21%, which depicts that BRICS banks are higher capitalized over European ¹⁸, Asian ^{2, 47} and developing economies ⁵. The average values of other variables are also presenting in the 2^nd column of the table.

Pearson's correlation coefficients matrix is presented in Table 4. In Table 4, we observe no coefficient value possess a high degree of correlation between the independent variables, so it can be assumed that our models are free from significant multicollinearity problems.

Two alternative measures of bank efficiency are used to examine the effect of competition and development indicators. Table 5 - Table 6 portrays the GMM estimators of Equation (1) by applying an alternative measure of competition LI, HHIA, and HHIGL. Table 5 - Table 6 presents the effect of competition and development indicators over the efficiency of cost and revenue, respectively. The coefficient of lagged dependent variable of efficiency is significant, which implies the dynamic nature of the model and persistently determined from one year to year (Table 5 - Table 6). In Table 5(6), we observe the coefficients of LI, HHIA, and HHIGL are significant and negatively related to the efficiency of cost (revenue), indicating a negative impact of market power over the efficiency of BRICS banks. The coefficient of HHIA and HHIGL refers that both asset and loan concentration decreases the efficiency of BRICS banks. That means increased market competition leads the BRICS banks toward revenue and cost-efficiency. These results are in line with the finding of ², ¹⁰, and ⁶ and also reinforce the 'quiet life' hypothesis on BRICS countries.

The coefficient of industry-level control variable BSD shows the positive (negative) association with the efficiency of cost (revenue) in Table 5(6). It signifies that with the banking sector development, the efficiency of cost (revenue) increases (decreases). These results support the previous study of ² and ⁶. The macro-economic variable-inflation shows no significant relationship with banks' efficiency, whereas economic progression-GDP shows a significant negative (positive) association with the efficiency of cost (revenue). One possible reason for such association is that with economic progression, banks' overall loan exposure increases, positively affecting their revenue efficiency and inversely affects the cost efficiency due to low monitoring of loans and advance and proportionate growth of non-performing loans.

Moreover, additional fund demand may stimulate banks to aggressive banking that increases the cost of the fund. These results are analogues with the empirical studies of ⁴. With the increase of assets exposure, i.e., banks' size efficiency also increases, denoted by the significant positive coefficient of size. In explaining other bank-level control variables in the efficiency of the cost model, we observe that capitalization (cap to TA) and profitability (ROA) enhance the cost efficiency of banks. It means that capitalized banks are more cost-efficient than low capitalized counterparts. Again performance through return on assets also amplifies the efficiency of banks.

Like the previous study of ^{39, 48, 49, 50, 51}, the finding of this study also spectacles a negative association of risk and efficiency (see Table 6) of banks. ^{39, 49} support the moral hazard hypothesis^{2} against the negative relationship between efficiency and risk. Again, ⁵⁰ opine that banks' risk is subject to low cost and revenue efficiency. The negative coefficient of ROA in the revenue model's efficiency depicts banks having more profit but are gaining less revenue efficiency than other counterparts. Diversified sources of income act positively in banks' efficiency, denoted by RD's significant positive coefficient in the efficiency of revenue model ³⁰.

Following the study of ^{2, 9, 10}, we also use square terms of market competition and development indicators to delve into the nonlinear relationship between competition, development indicators, and banks' efficiency. Table 7 - Table 8 presents the GMM estimators having a nonlinear impact using Equation (2).

In examining the size and competition impact, we observe that size and market competition have a homogeneous effect on the efficiency of cost. The interim term Small × Com (Large × Com) coefficients in all inverse competition measures (HHIA and HHIGL measures) have found positive and significant. It refers to the increase in market competition; the cost efficiency of BRICS banks decreases irrespective of size. However, in the long run, the efficiency increases, denoted by the size × competition square term (a square measure of competition) at HHIA in Model I and HHIA & HHIGL in Model II. The efficiency of cost behaves just in the opposite pattern of competition in the economic progression. In the economic progress, i.e., with the increase of GDP, the cost efficiency of BRICS banks initially increases (observed through the interim term of size and GDP) and subsequently decreases (observed through square term of size and GDP) at Model I-II in HHIA and HHIGL measures.

The revenue efficiency measure discloses some interesting facts of BRICS banks (see Table 8). Market competition and development indicators (GDP) have a meaningful association with small banks' revenue efficiency. Large banks have shown a significant association with revenue efficiency in LI measure. However, no significant association is observed between GDP and revenue efficiency of large banks. Small banks' revenue efficiency of BRICS nations in competitive market situation initially increase and decrease in the long run where large banks also show the same trend in the short-run. With the GDP growth, small banks' revenue efficiency initially decreases then increases in the long run.

Thus, competition and development indicators have a homogeneous effect on different size of BRICS banks in each efficiency measures; however, their effect is heterogeneous when the efficiency measure is different.

Few studies address that competition influences the performance, measured through efficiency ^{6, 9}. We extend their works by introducing size, competition, and development indicators effect through linear and quadratic examination. Moreover, this study covers the GDP of examining joint effect competition and development indicators in different BRICS banks' efficiency exposure.

6. Concluding Remarks

Bank, the well-known dominant and significant player of the financial system, actively plays a role in circulating economic flow and economic progression. Hence, it is worth considering the factors that affect the efficiency of banks.

Our study differentiates from existing literature in several aspects. Firstly, most of the works are done by examining the competition, performance & inefficiency of banks based on single countries ⁴. So we fill up the gap of existing literature by introducing the effect of size, competition & development indicators in the efficiency of BRICS banks. Secondly, this study is original in the sense that it shows the impact of competition & development indicators on the efficiency of revenue & cost by using the two-step system GMM of BRICS banks. In analyzing descriptive statistics, it is found that market competition of BRICS countries banks is increasing day by day as the mean of Linder Index (LI) & Hirschman Index (HHI) of latest 3 years are getting less than the overall mean value of all concentration in all countries. Aftermaths of baseline and extended regression results, it is observed that bank competition & development indicators substantially impact the efficiency in the region. The main results find that the coefficient of LI, HHIA & HHIGL, i.e., market power have a negative impact over the efficiency of BRICS banks. It indicates that the increase of market competition leads the BRICS banks to revenue & cost efficiency, which confirms the "Competition–efficiency" hypothesis.

The findings also suggest that there is a nonlinear relationship between competition, development indicators, and efficiency. Extended results also depict that competition has a similar effect on the different size of banks on each type of efficiency; however, the competition and size heterogeneously affect banks' efficiency. Moreover, the quadratic effect of competition also found heterogeneity on the efficiency of BRICS banks.

In economic progression, the revenue and cost efficiency also behave differently. With economic progress, the cost efficiency decreases and revenue efficiency increases. By examining the size & competition impact, it is shown that size and market competition have a homogeneous effect on the efficiency of cost and efficiency of revenue. However, the results are heterogeneous in different efficiency measures. The nonlinear relationship also suggests that small and large banks become more revenue efficient initially in a competitive market than in the long run. Thus it can be said that competition & development indicators heterogeneously affect the efficiency of banks. Finally, further study can be extended by focusing on developed regions on the tilted study. Other efficiency measures like human capital efficiency and stability efficiency can also be absorbed to extend further research.

Notes

1. The "quiet life hypothesis (QLH)" postulates that due to weak monitoring, managers enjoy the market power at the cost of low level of profit and cost saving as they are reluctant to take risky decision.

2. Moral hazard hypothesis holds that low capital ratio induces banks to take riskier project resulting increased credit risk in future.

References

Appendix A

Determination of Cost and Revenue Efficiency Using Stochastic Frontier Analysis (SFA).

The stochastic frontier analysis originated by ⁵² is used to calculate each bank's efficiency based on the production frontier. On this production frontier model, the stochastic cost frontier model was developed (For details, see ^{39, 53}). According to this methodology, due to inefficiency and random noise, a bank's observed cost is formulated to deviate from the cost-efficient frontier ⁵⁴.

For the nth Bank,

(1)

Where TC_n represents total operating cost, including financial costs, Q_i indicates three outputs, i.e. Q₁=Gross loans and advances, Q₂= Deposit and short-term funding, Q₃= Total security. Pj stands for two input prices, i.e. P₁= price of the fund, which is the ratio of total interest expenses to total deposit, P₂= Price of physical capital, which is non-interest expenses to fixed assets. ε_n shows the deviation of the actual total cost of a bank from the cost-efficient frontier, and it has two disturbance terms given as below:

Where V_n is the random error term, and we assume that this is independent and identically distributed N (0,). U_n represents cost inefficiency and assumed to be distributed independently of V_n and a half-normal distribution, i.e. N (0,).

By using the intermediation approach ⁵⁵ and by following ⁵⁴, we have developed the following multiproduct translog cost function to specify the cost function:

(2)

According to ⁵⁶, the expected value of U_n, on conditional ε_n, represents the cost-inefficiency of bank n (which is defined as C_n).

(3)

Where λ is the ratio of the standard deviation of U_n to standard deviation of V_n, φ is the cumulative standard normal density function, and ϕ is the standard normal density function. C_n can be estimated by using Equation (3).

We also use the alternative revenue efficiency specification, where the dependent variable is Rev_n = Pre-tax profit of all banks in the sample. Inputs and outputs are the same as used in cost efficiency. The composite error term is now defined as V_n– U_n. The general procedure for estimating production inefficiency in Equation (3) is to estimate coefficients and the error term ε_n= V_n– U_n first, and then calculate the efficiency for each observation in the sample. We just alter the error term to V_n– U_n from V_n + U_n to use the Equation as a production function ⁵⁷. And here U_n represents human capital inefficiency and assumed to be distributed independently of V_n and a half-normal distribution, i.e. N (0,). We use computer software called Frontier Version 4.1c developed by, ⁵⁷ for Stochastic Frontier Production and Cost function estimated by the method of maximum likelihood.

Appendix B