MEASURING COMPETITION IN LESOTHO BANKING SECTOR
A Thesis Submitted to the Department of Economics,
FACULTY OF SOCIAL SCIENCES,
THE NATIONAL UNIVERSITY OF LESOTHO
in partial fulfilment of the requirements for the degree of
Master of Arts in Economics
LIST OF TABLES
Table 1.1Banks’ Market share of assets by ownership (%)
Table 4.1Summary statistics
Table 4.2Correlation Analysis
Table 5.1Long-run equilibrium test
Table 5.2Regression Results
LIST OF FIGURES
Figure 4.1Unit prices of deposits and labour
Figure 4.2Unit price of capital
Figure 5.1Herfindahl-Hirschman Index
Figure A1Interest rate spread
Figure A2Composition of banks revenue
TABLE OF CONTENTS
TABLE OF CONTENTS
TOC o “1-3” h z u 1INTRODUCTION PAGEREF _Toc512227865 h 11.3The Lesotho Banking Sector PAGEREF _Toc512227866 h 22Literature Review PAGEREF _Toc512227867 h 52.1Theoretical Literature PAGEREF _Toc512227868 h 52.2Empirical Review PAGEREF _Toc512227869 h 73Methodology PAGEREF _Toc512227870 h 113.1Theoretical Model PAGEREF _Toc512227871 h 113.2Empirical Methodology PAGEREF _Toc512227872 h 124Data and Summary Statistics PAGEREF _Toc512227873 h 144.1Data PAGEREF _Toc512227874 h 144.2Correlation Analysis PAGEREF _Toc512227875 h 195Discussion of Results PAGEREF _Toc512227876 h 205.1Concentration Analysis PAGEREF _Toc512227877 h 205.2Long-Run Equilibrium Test PAGEREF _Toc512227878 h 215.3Regression Results PAGEREF _Toc512227879 h 226CONCLUSIONS AND RECOMMENDATIONS PAGEREF _Toc512227880 h 24REFERENCES PAGEREF _Toc512227881 h 26APPENDIX A PAGEREF _Toc512227882 h 29
1INTRODUCTIONA healthy competition in the financial sector is of utmost importance for several reasons. Like in many other industries, the extent of competition could have implications on efficiency and quality in the provision of financial services, as well as innovation in the financial sector, (Claessens, 2009). This is because when there is adequate competition, banks would seek the best methods of utilizing their resources in such a way that minimizes costs but generates higher profits, in order to stay competitive in the market. Efficiency in the banking industry could ultimately lead to provision of financial services at lower costs. For instance, businesses and households would obtain funds from the banks at lower costs, thus leading to more private sector investment that could stimulate economic growth. The Lesotho’s banking industry is characterised by few number of banks and high levels of concentration. The high degree of concentration raises fears about the competitive conditions in the industry. Although highly concentrated banking markets raise worries about adequacy of competition, previous studies in banking industries show that banks can still act competitively even in highly concentrated banking industries, (Shaffer, 1993; Baumol et al., 1982). The theory of contestable markets proposed by (Baumol et al., 1982) argues that concentrated markets can still be competitive if there are low barriers of entry and exit. Hence, one cannot conclude about the competitive conditions of the industry by just looking at concentration indices.
This thesis addresses the following questions: firstly, what is the degree and nature of competition in the banking sector in Lesotho? Secondly, how does the competitive conditions of the banking sector in Lesotho compares with other African banking markets?
This study contributes in many ways to the literature on banking sector competition. Although there is a wide research on the topic of measuring the degree of competition in the banking sector, only few studies cover this topic in the context of African banking markets. To be precise, to the best of the researcher’s knowledge, there is no study that analyses the nature and degree of competition in the Lesotho’s banking industry. The findings of this thesis will be of particular importance to policy makers, academics and the consumers of banking services in Lesotho. Knowledge of competitive conditions in the banking sector will assist policy makers to amend and or introduce policies aimed at stimulating competition in Lesotho’s banking industry.
The rest of this thesis is organized as follows: The next section provides a review of the Lesotho’s banking industry. It provides a detailed explanation of how the formal banking industry began, how it changed until now, as well as its relationship with the South African banking industry. Chapter 2 reviews key theoretical and empirical studies that evaluate competition in the financial sector. Chapter 3 outlines the methodology that is used to estimate the degree of competition in Lesotho’s banking industry. Chapter 5 presents analysis of results, while chapter 6 concludes by providing a summary of the thesis and its recommendations.
1.3The Lesotho Banking SectorThe formal banking in Lesotho dates back to early 1900’s, with the arrival of Standard Bank, a British Bank, which started its operations in 1902 (Maruping 1992). Participation of foreign banks increased when the second foreign bank, Barclays Bank, came to Lesotho in 1957. Barclays Bank focused more on corporate clients than individuals. In 1977, Standard Bank merged with the Chartered Bank to form a new bank, the Standard Chartered Bank. The merger might have hampered competition in one way or the other given the limited number of banks that were operating in Lesotho.
The government of Lesotho established the Lesotho Bank in 1972. This was because the commercial banks were channelling funds from Lesotho to South African markets, thereby financing South Africa’s development projects. Another reason for founding of the Lesotho Bank was that commercial banks were paying customers very low interest rates on their deposits. The authorities initially anticipated the bank to serve as both a commercial bank and a development bank. However, in a short while the bank served purely as a commercial bank. Lesotho Bank operated as commercial bank, although it was envisaged to be a development bank. The bank served mostly businesses owned by Basotho and foreign owned companies operating in Lesotho. In mid-1980’s, Lesotho bank was the largest of the commercial banks operating in Lesotho. The bank provided services to Basotho owned businesses as well as foreign owned businesses operating in Lesotho. During that period, Lesotho Bank was the largest of the commercial banks operating in Lesotho.
Participation of state owned banks increased in 1975 when the government established the Lesotho Agricultural Development Bank (LADB). The bank commenced its operations in 1976. As a result, there were three commercial banks and one development bank in operation in Lesotho. The main reason for the establishment of the LADB was to provide financing for Agriculture. In its first ten years of existence, the LADB extended credit, financed mainly through grants as well as concessional loans from international organizations. The LADB in late 1980s increased its branch network to various areas where the bank did not reach previously. The bank’s increased capacity was however not been fully utilized given that some of the areas that the bank increased its braches to, were less populated. This rapidly increased the bank’s operating costs relative to its income. This challenge, together with increasing defaults by the bank’s customers on loans granted resulted in the LADB being closed in 1998.
The banking market continued to experience major structural changes from mid-1990s to late 1990s. In 1995, a new foreign owned bank, Nedbank Lesotho joined the banking industry. It took over the operations of Standard Chartered Bank. The bank provided services mainly to corporate, business and VIP customers. In a similar year (1995), Standard Bank took over the Barclays Bank. The Lesotho Bank was partially liquidated in 1999 following poor financial performance of the bank. Subsequent to partial liquidation of Lesotho Bank in 1999, Standard Bank purchased a 70 percent share in the bank. These developments reduced the dominance of government-owned banks to private-foreign owned banks.
The Lesotho’s banking system is dominated by foreign-owned banks. As of December 2016, the banking industry consisted of four commercial banks, with 49 branches. Of the four banks, three are subsidiaries of South African banks. The three foreign banks account for about 97 per cent of total banking sector’s assets. The fourth bank (Lesotho Post Bank) is the only domestic bank and is fully owned by the government of Lesotho. Table 1.1 shows the market share of foreign banks compared to that of local-owned banks.
Table 1.1Banks’ Market Share of Assets by Ownership (%)
2013 2014 2015 2016
Foreign Banks 97.26 97.17 96.81 93.87
Domestic Banks 2.74 2.83 3.19 6.13
Total 100 100 100 100
Table 1.1 depicts market share of foreign vis-a-vi local banks in terms of their total assets. Source: Author’s own computations based on CBL data.
The banks seem to be working towards improving their distribution channels and also implementation of modern technologies so as to increase the reach of financial services that they offer. For example, number of automated teller machines (ATMs) has increased by about 48 per cent from 138 in 2013 to 204 in 2016. In a similar manner, Point of Sale machines increased significantly from 802 in 2013 to 1374 in 2016. This represents an increase of more than 70 percent in three years. The rapid improvement of the distribution channels might be a resultant of a fast growing mobile money services (M-PESA and Eco-cash) that are offered by telecommunications companies in Lesotho. With respect to commercial banks number of branches, there has only been addition of two branches between 2013 and 2016. Table 1.2 shows the commercial banks distribution channels in terms of number of banks, commercial banks branches, ATMs as well as POS machines.
Table 1.2Commercial Banks Distribution Channels LINK Excel.Sheet.12 C:\Users\kthokoa\Desktop\March_2018_macroeconomic_data_Annual_Final.XLSX “March 2018!R35C13:R40C17” a f 4 h * MERGEFORMAT
2013 2014 2015 2016
Number of Banks 4 4 4 4
Number of Branches 45 45 47 49
Number of ATMs 138 159 180 204
Number of POS Machines 802 1050 1168 1374
Source: Central Bank of Lesotho
Close to rsa….move funds to rsa…crdit short…
The banking industry, for the past ten years has remained profitable, and maintained good quality assets. Figure 1.1 depicts measures of commercial banks profitability as measured by return on assets (ROA) and return on equity (ROE) from the year 2006 to 2015. The return on assets ratio measures the ability of banks to generate profit out of assets that they own. On the other hand, the return on equity reveals how effective do banks use their investors’ money. The ROA averaged 2.8 per cent while the average ROE registered an average of 29 per cent from 2006 to 2015. This indicates that banks have been using their assets and their owners’ equity efficiently to generate profit for the past ten years.
Figure 1.1Commercial Banks’ Profitability
Figure 1.1 shows commercial banks’ profitability ratios, return on assets (ROA) and return on equity (ROE) from 2006 to 2015. Source: Author’s own computations based on CBL data.
Figure 1.2Non-Performing Loans to Total Assets
Table 1.2 depicts the banking industry ratio of non-performing loans to total asset. Author’s own computations based on CBL data.
2Literature Review2.1Theoretical LiteratureThe literature on measurement of competition in banking industry can be broadly categorised into two main streams, being the structural and non-structural approaches. The structural methods emanate from the structure-conduct-performance (SCP) hypothesis. The SCP hypothesis, first proposed by Mason (1939) and further developed by Bain (1951) argues that banks operating in markets that are more concentrated tend to act collusively. This makes colluding banks to earn higher than normal profits because of uncompetitive practices such as fixing prices and limiting output.
The SCP hypothesis has however been criticised by a number of authors (Demsetz, 1973; Baumol et al., 1982) due to its lack of theoretical foundations and other practical drawbacks. According to Demsertz (1973), efficiency is the main determinant of a firm’s profitability as opposed to high degree of concentration put forth by the SCP hypothesis. The efficiency hypothesis argues that firms that are more efficient are able to grow faster, hence leading them to acquiring a large market share than less efficient firms. As firms that are more efficient grow, the efficient hypothesis claims that less efficient firms have a tendency of going out of the market.
The limitations of structural models let to the emergence of non-structural models that fall under the New Empirical Industrial Organisation (NEIO). The NEIO approaches test competition and do not make use of structural measures, such as concentration ratios when analysing competition among banks, (Bikker and Haaf 2002). The major advantage of NEIO measures of competition over traditional structural methods is that the NEIO approaches are derived formally from economic theory of firm’s profit maximization conditions. There are two non-structural models that have received much empirical attention in the banking industry; the Bresnahan’s market power and the Panzar-Rosse’s, models.
The Bresnahan (1982) and Lau (1982) model, later developed by Bresnahan (1989) makes use of general equilibrium conditions to analyse the degree of competition in the financial sector. The model is based on the premise that in equilibrium, profit-maximizing firms choose prices and output in such a way that marginal costs equal to perceived marginal revenue. By estimating simultaneous equation model consisting of the demand relation and first order condition of firm’s profit maximization, an index or parameter is estimated which provides a measure of the degree of market power exercised by firms. The key advantage of this model is that it can even be employed using only aggregate sector data rather than firm-specific data.
The second non-structural model commonly applied in empirical studies is the Panzar-Rosse (1987). It shows how revenues react to changes in banks prices of inputs. The Panzar-Rosse model uses the measure known as the H-statistic to assess the level of competition in the industry. In addition, the H-statistic is used to analyse the underlying market structure of an industry, (Mirza et al 2016). The H-statistic is given as the sum of elasticities of banks revenue to changes in input prices. When the H-statistic is less than or equal to zero, the market structure is characterized as monopoly or perfectly colluding oligopoly, (Panzar and Rosse, 1987). Monopolistic competition is supported for H-statistic greater than zero but less than one. Lastly, perfect competition is consistent with the H-statistic value of one, which is the upper limiting value of the statistic.
2.2Empirical ReviewSeveral empirical studies have been conducted using various techniques to analyse the competitiveness of the banking sector in different countries. Some studies do cross country comparisons while others are based on single markets. Bikker and Haaf (2002) assessed the degree of competition and examined the impact of concentration on competition in European banking markets. The study employed Panzar-Rosse methodology and individual banking data ranging from the period 1988 to 1998 to measure the extent of competition. Their results reveal that banking markets in the industrial world are largely characterised by monopolistic competition. They further found that on average large banks are more competitive compared to smaller banks.
Claessens and Laeven (2004) used the methodology of Panzar-Rosse (1987) in fifty countries, to assess the extent to which changes in input prices are reflected on banks revenues. The study used panel dataset spanning from 1994-2001 to estimate the reduced form revenue equation using both OLS with time dummies and GLS with fixed bank effects. The H-statistic mostly varied from 0.6 to 0.8, indicating that the banking sectors of the studied countries are monopolistically competitive. In addition, the study suggests that participation of more foreign banks and fewer restrictions on banks activities boosts competition among banks. This is largely consistent with Baumol et al., (1982) argument that low entry and exit barriers makes incumbent banks to be competitive due to threat to entry of new firms.
By using the Panzar-Rosse methodology, Aktan and Masood (2010) analysed the competitive structures of the Turkish banking industry from 1998 to 2008 for 17 large banks. The paper explored the factors that explain differences in the degree of competitiveness among banks. The results of the long-run equilibrium test showed that the Turkey banking was in long-run equilibrium state. Hence, the use of Panzar-Rosse method was valid. The H-statistic was 0.753, which means that Turkish banks were operating under conditions of monopolistic competition. All input prices appeared with the positive sign. The price of funds, measured by the ratio of interest expenses to total deposits, was the main contributor to the value of H-statistic, with a coefficient of 0.331.
Karasulu et al. (2007) analyzed the banking market competition in Chile, comparing it with it with that of 28 other emerging market countries by applying the Panzar and Rosse’s methodology to bank level data for 1995-2004. The estimation results suggests that Chile’s banking industry like most banking systems can be classified under monopolistic competition. However, the Chile’s banking system had a relatively lower competition index compared to cross-country sample. Moreover, the researchers found that Chilean banks were more profitable than other emerging markets banks. This is not surprising due to the fact that Chile’s banking market was found to be less competitive, coupled with the fact that on average, net interest margin of Chilean banks was found to be between 1.50 and 2.25 points higher than the average bank in the sample, after controlling for macroeconomic and sector-specific factors that may affect profitability. The lower competition was attributed to lower elasticity of revenue to funding costs as well as weak competition from nonbank financial institutions.
The study in Brazil banking sector by Barbosa et al. (2014) investigates the competitive aspects of multi product banking operations. Freixas et al., (2007) define multiproduct banks as entities that offer financial, banking and insurance services using the same corporative structure. Barbosa et al. (2014) study extends the Panzar-Rosse (1987) model by developing a theoretical model that helps to study the difference in behaviour and conduct of banks that offer only banking services and those that offer additional services such as insurance services. They found that banks that provide only additional services have higher market power than banks that offer only typical banking products. This is largely explained by the finding that banks that are more diversified in their product offering benefit from economics of scope.
The study found that economies of scope helps in reducing marginal costs for multi-product banks for each product offered. Hence, these types of banks are able to set higher margins over their marginal costs in both banking and non-banking products compared to banks that provide only the normal banking services. Moreover, the findings of this study suggest ignoring the multi-product nature of banks may lead to overestimation of the Panzar-Rosse H-statistic. This means that the level of competition is overestimated and this could lead to misleading policy recommendations. The Barbosa et al. (2014) approach is more applicable in advanced banking markets where banks offer additional services such as insurance, life insurance and bonds.
There have also been a number of studies conducted in African countries that assess competitiveness of the banking sector. Fosu, (2013) examined competition across sub regional banking markets in Africa. The study employed the Panzar-Rosse model to assess the overall degree of banking competition in every sub regional banking market between 2002 and 2009. The results reveal that generally, African banking markets exhibit monopolistic competition behaviour as in many emerging market economies. The results further indicate that recent structural reforms in Africa might have contributed positively to enhancing competition in African banking sectors.
One of the recent, single country studies in Africa which also uses the Panzar-Rosse methodology was conducted by Simatele (2015) and examines the relationship between market structure and competition in South African banking sector. In order to investigate how competition has evolved over the sample period (1997-2014) the paper makes use of a time-dependent Panzar-Rosse H-statistic. The study found the South African banking sector to be highly concentrated, with the four largest banks accounting for more than 80 percent of the total market share. In addition, the results reveal that the banks in South Africa operate in a monopolistically competitive market, with the H-statistic varying from 0.431 to 0.711 depending on the dependent variable used. The time varying H-statistic suggested that competition in the industry has increased over the sample period, although by small amount. The results are consistent with the findings of Simbanegavi et al (2014) who applied the Bresnahan and Panzar-Rosse model in South African banking industry.
Hamza (2011) examined the level of competition in Tunisian banking sector using the Panzar-Rosse (1987) model. The study used a dataset covering the period from 1998 to 2008. The results suggest that monopolistic competition is the prevailing market structure in the Tunisian banking sector. The results verified the finding of previous studies that suggest that there has been a transition from monopoly structure to monopolistic competition. This transition is said to be caused by the efforts of Tunisian authorities to improve competition in the banking sector.
3MethodologyThis chapter outlines the methodology used to assess the state of competition in the Lesotho’s banking industry. The first section of the chapter begins by presenting the theoretical methodology of the Panzar-Rosse methodology. The second section outlines the empirical methodology used in the thesis.
3.1Theoretical ModelThe derivation of Pazar-Rosse model is based on economic theory that the objective of a firm is profit maximization. The profit function of a profit-maximizing bank is given as:
?=Riyi, n, Zi-Ciyi, wi, Xi (3.1)
Where ? denotes profit of the ith bank, while Ri and Ci are the revenue and cost functions of the ith bank respectively. The variable yi stands for bank i’s output level, n represents the number of banks operating in the market, wi denotes the vector of input price of bank i while Zi and Xi stand for vector of exogenous variables that could shift bank the ith bank’s revenue and cost functions, respectively.
Taking the first derivative of the profit equation to derive the profit maximizing condition for bank i, yields:
Ri’yi, n, Zi=Ci’yi, wi, Xi (3.2)
Where Ri’ and Ci’ represents the marginal revenue and marginal cost of the ith bank respectively. The profit maximising output is therefore given as y*=yi*(Zi,wi, Xi). Hence, the total revenue equation becomes:
Ri*=Ri*yi*(n*, Zi,wi,Xi) (3.3)
The variables with asterisk denote equilibrium values. The H-statistic, which is the sum of elasticities of the reduced form revenue function with respect to input factor prices is:
H= k=1K?Ri*?wkiwkiRi* , (3.4)
where wki is factor k input price for the ith bank.
As indicated by Simatele (2015), the dependent variable used depends on whether the researcher assumes the main function of the commercial banks is intermediation or if other revenue sources are as important as well. It is suggested that interest revenue be used in the first case and the sum of interest and non interest revenue be employed in the latter case. Hence considering that a significant portion of banks revenue in Lesotho is made up of interest revenue, it is assumed that the core business of the banks is intermediation.
3.2Empirical MethodologyThis section outlines the empirical methodology used to assess the state of competition in the Lesotho’s banking industry. This study adopts the method employed by Bikker et al. (2012), however with more control variables that capture for bank specific factors that affect the dependent variable. This thesis uses the following empirical version of the reduced form revenue equation for the purpose of finding the H-statistic and hence analysing the degree of competition in banking industry in Lesotho:
ln INTREVit=?0+ ?1ln LPit+ ?2ln DPit+?3ln CPit+?4lnLLPit+?5lnLTAit+?6lnEQTYit+?7ln OIit+? (3.5) where:
– the subscripts i and t denote bank i and month i, respectively;
– ln INTREV represents total interest income, which is a proxy for banks’ income;
– lnLP is the ratio staff costs to total assets, a proxy for unit price of labour;
– ln DP is the ratio of interest expenses to total deposits and is used to approximate the total cost of deposits;
– ln CP is the ratio of operating expenses to total fixed assets – a proxy for price of capital;
– ln LLP, ln LTA and ln EQTY denote the ratio of non-performing loans to total loans, loans to total assets and equity to total assets respectively. These factors are included to reflect variations in risk, cost, size and structure of banks;
– ln OI represents the ratio of other income to total assets, and is intended to see the effect of changes in bank’s other income on interest revenue;
– ? is the random error term.
We estimate equation (3.5) using the Feasible Generalised Least Squares method (as suggested by Bikker et al. 2012) to cater for the problem of heteroscedasticity that may come because of using non-scaled revenue as the dependent variable. We calculate the H-statistic as the sum of elasticities of revenue to changes in cost of inputs, that is H=?1+?2+?3. We further perform the Wald test to find out if the obtained H-statistic is significantly different from zero or not.
For the Panzar-Rosse model to be implemented, it is required that the market be in long-run equilibrium. Buchs & Mathisen (2005) says that given the internal logic of the Panzar-Rosse model, long-run equilibrium reflects the banks’ ability to adjust to shocks. Following literature, we test this assumption by estimating the model outlined in equation (3.5) but replacing the dependent variable by the return on assets (ROA). Return on assets is the ratio of net income to total assets. The estimated model is as follows:
ln (1+ROAit)=?0+ ?1ln LPit+ ?2ln DPit+?3ln CPit+?4lnLLPit+?5lnLTAit+?6lnEQTYit+?7ln OIit+? (3.6)
Since ROA has a potential of taking on small negative values hence making it impossible to take the natural logarithm, most studies use the modified dependent variable asln (1+ROAit), (Claessens and Laeven 2004). The H-statistic based on the return on assets (HROA) is calculated and is defined as ?1+?2+?3. The long-run equilibrium condition is satisfied when (HROA=0). We also perform the Wald test to find evidence of whether the H-statistic based on return on assets is different from zero. Shaffer (1982) says that the intuition underlying this long-run equilibrium test is that at equilibrium banks return on assets should not be associated with banks input factor prices.
4Data and Summary Statistics4.1DataThe study uses balanced monthly panel data set for all the four commercial banks operating in Lesotho for the period 2013:9 to 2016:12. The data was obtained from the Central Bank of Lesotho. The choice of period for analysis was based mainly due to the availability of data. The years prior to 2013 had too many missing data points, which could then undermine the credibility of the results of the study.
Table 4.1 reports the summary statistics of the variables used in this thesis. The data is recorded in thousands of maloti for easy of handling. The statistics reported are the mean, standard deviation, minimum and maximum of all the variables used. The standard deviations of the variables include the between and within observations standard deviation in addition to the overall standard deviation. The between standard deviation is used to capture the variability between the banks. The within standard deviation on the other hand is included in order to capture the variance within each bank over the study period.
The average interest revenue on a monthly basis, for the study period is about M24, 670.00. The minimum of which is M1, 161.00 while the maximum is M75, 812.00. There is huge variability between the banks’ interest revenue. The standard deviation of interest revenue between banks is about 27000 while in within banks is 4123. Overall dispersion in interest revenue is 23817. Large variability of interest revenue between banks means that the revenues are not evenly spread among banks. In a competitive market, firms’ market shares are evenly distributed, so that no firm has market power. Huge variance of revenue suggests that some banks are highly dominant over others and hence have a large degree of market power.
Looking at the measures of input prices, the average of cost of capital is the highest. The mean of capital price is about 0.21, compared with the unit cost of deposits and labour of about 0.003 each. The large costs of capital reflect high banks operating costs. As with interest revenue, there is high variability between banks’ factor prices than the variability within banks. Considering control variables, loans to total assets and operating income have the highest averages; 0.35 and 0.37 respectively.
Table 4.1 Summary statistics
Variable Mean Standard Deviation Minimum Maximum Observations
Interest Revenue overall 24670.57 23817.46 1161 75812.06 N = 156
between 26999.73 3254.56 63462.81 n = 4
within 4123.713 11696.22 37019.82 T = 39
Deposits Price overall 0.002962 0.00165 0.000421 0.009689 N = 156
between 0.001751 0.000739 0.005022 n = 4
within 0.000638 0.001289 0.007629 T = 39
Labour Price overall 0.003068 0.001537 0.0008 09 0.006579 N = 156
between 0.001645 0.001528 0.005213 n = 4
within 0.000567 0.000481 0.005187 T = 39
Capital Price overall 0.21095 0.178043 0.008605 0.93278 N = 156
between 0.192773 0.020318 0.40462 n = 4
within 0.060405 0.035862 0.79713 T = 39
Equity to total assets overall 0.113615 0.034038 0.079183 0.250423 N = 156
between 0.020975 0.100338 0.144805 n = 4
within 0.028749 0.06837 0.219233 T = 39
Loans to total assets overall 0.353617 0.09569 0.041084 0.491446 N = 156
between 0.08692 0.246729 0.439197 n = 4
within 0.05877 0.073378 0.466117 T = 39
Loan-Loss Provision overall 0.0474237 0.058323 0.083135 0.525311 N = 156
between 0.017708 0.034803 0.072929 n = 4
within 0.056257 0.017192 0.499806 T = 39
Operating Income overall 0.3666675 0.119914 0.123169 0.633699 N = 156
between 0.127698 0.275445 0.554566 n = 4
within 0.045502 0.189115 0.505908 T = 39
Notes: Table 4.1 shows the descriptive statistics of the variables used in the Panzar-Rosse model. Data recorded in millions of Maloti. Source: Authors own computations based on commercial banks data from CBL.
Figure 4.1 below depicts trends of unit costs of deposits and labour. The price of deposits is measured as a percentage of interest expenses to total deposits. Interest expense is the price banks pay to their customers for the deposits they keep at the banks. The labour costs are measured by ratio of total labour costs to total assets, expressed as a percent. The unit costs of deposits have been fluctuating around 0.4 percent from late 2013 until beginning of 2016. Thereafter, the unit costs of deposits have shown an upward trend. The unit costs of labour, as shown by the dashed line, have been rather constant from 2013 but the costs seem to rise since early 2016.
Figure 4.1Unit Price of Deposits and Labour
Notes: Figure 4.1 depicts unit prices of deposits and labour for the entire banking sector between October 2013 and December 2016. The ratio of interest expenses expressed as a percentage of total deposits are used as a proxy for cost of deposits while labour costs expressed as a percentage of total assets is an estimate of unit price labour. Source: Authors own computations based on commercial banks data from CBL.
Figure 4.2 shows the unit price of capital. The price of capital is measured as a ratio of operating expenses to total fixed assets. Operating expenses are expenses incurred by bans through their usual operations. The average cost of capital was about 25 percent in 2013. In the subsequent years, the unit price of capital declined and reached a low 17 percent. The high levels of average cost of capital suggest that banks experience high operating expenses. High costs of capital may have adverse impact on customers. These come in the form of high banking services fees as banks try to cover up the operating expenses. Again, the large operating expenses may have some implications on the efficiency of banks. That is the banks’ ability to turn resources into revenue.
Figure 4.2Unit Price of Capital
Notes: Figure 4.2 above displays the ratio of operating expenses to total fixed assets expressed as a percentage. This is used to approximate the unit price of capital for the commercial banks between 2013 and 2016.
4.2Correlation AnalysisTable 4.2 below is the correlation matrix, which depicts the linear relationships between all the variables used in this study. The correlation coefficients only show the direction and strength of the association between the two variables without necessarily implying that one causes the other. The correlation coefficient ranges from -1 to 1. The elements in the leading diagonal of table 3.1 show the correlation between each variable and itself. All the correlations in this diagonal are equal to one. This means that there is a perfect positive correlation between a variable and itself. There is a strong correlation between input prices and the interest revenue. The correlation between price of deposits and interest revenue is about 0.31 implying a moderate relationship. Prices of labour and capital reflect a negative relationship with interest revenue, with correlations of about -0.59 and -0.76 respectively. The correlations among independent variables fall within the recommended limit of 0.8. Hence, this shows absence of perfect linear association among the explanatory variables. The largest correlation between two independent variables is that between deposits price and labour price, with a correlation coefficient of about -0.79.
Table 4.2Correlation matrix
Interest Revenue Deposits Price Labour Price Capital Price Equity to total assets Loans to total assets Loan-Loss Provision Operating Income
Interest Revenue 1 Deposits Price 0.3147 1 Labour Price -0.5862 -0.7852 1 Capital Price -0.7588 -0.4822 0.6773 1 Equity to total assets -0.189 -0.4141 0.4123 0.2392 1 Loans to total assets 0.3941 -0.2253 0.1495 0.095 0.1058 1 Loan-Loss Provision -0.1584 -0.1759 0.1151 0.046 -0.1184 -0.5729 1 Operating Income -0.4136 -0.1738 0.312 0.6663 -0.1953 0.3034 0.0024 1
Notes: Table 4.2 is the correlation matrix showing the degree of association between variables. Source: Authors own computations based on commercial banks data from CBL.
5Discussion of Results5.1Concentration AnalysisThe study uses Herfindahl-Hirschman index (HHI) to assess the degree of concentration in Lesotho banking industry. However, the concentration index is meant to only give a picture of the level of concentration in banking industry but not to infer the degree of competition or nature of the market structure. Figure 5.1 shows the trends in the HHI from 2013 to 2016. The index has been computed using total banks assets, deposits as well as loans using equation (1). The results indicate that the HHI on assets, deposits and loans have consistently been very high for the period under study, although there seems to be downward trend in the index. In particular, the HHI on loans is very high compared to that of deposits and assets. For example between 2013 and 2014, the HHI on loans was in excess of 5000 points while it was about 4500 for deposits and assets for the same time period. The market with HHI greater than 2500 is considered to be highly concentrated. Too high HHI values simply means that a large proportion of assets, deposits and loans is concentrated on only few banks. However empirical research has shown that a high concentration level does not necessarily imply low competition, Okeahalam (2002).
Figure 5.1Herfindahl-Hirschman Index
Notes: Figure 5.1 depicts the market concentration based on the Herfindahl-Hirschman index (HHI). The index is calculated by summing the squared market shares of (1) total deposits, (2) assets and (3) loans of individual commercial banks respectively. The HHI is for period from October 2013 to December 2016. Source: Authors own computations based on commercial banks data from CBL.
5.2Long-Run Equilibrium TestFor the estimation results of the Panzar-Rosse model to be valid, the banking industry should be in long run equilibrium. Table 5.1 shows the test results of equation (4.6) for assessing whether the observations are in long-run equilibrium. The null hypothesis is that HROA=0, against alternative hypothesis that it is statistically different from zero. The Wald test fails to reject the null hypothesis at 1, 5 and 10 percent level of significance. This leads to a conclusion that banking sector is in long-run equilibrium for the period under study.
Table 5.1 Long-run equilibrium test
ln(Deposits Price) -0.000416
ln(Labour Price) -0.000155
ln(Capital Price) -0.00134***
ln(Equity to Total Assets) 0.00102
ln(Loans to Total Assets) 0.000931
ln(Other Assets) -0.000778
ln(Operating Income) 0.00184**
Number of banks
Wald test: HROA=0 4
Notes: (***), (**) and (*) denote 1, 5 and 10 percent level of significance, respectively.
Standard errors are in parentheses
5.3Regression ResultsTable 5.2 reports the regression results of Panzar-Rosse model. From the estimation results, coefficient of deposits price is positive, as expected and statistically significant at all levels of significance. This means that an increase in deposits price by one percent, on average increases interest revenue by about 0.62 percent, all other factors remaining the same. This result is in line with many empirical studies in banking sector competition, (Simatele, 2015; Hamza, 2010 ; Mirza et al, 2016). Although it is generally expected that increases in expenses be negatively related to banks revenues, increases in deposits price has a positive net impact on banks interest revenue. This is because when price of loanable funds or deposits increase, revenue from loans (in the form of interest) also increases.
Unlike the coefficient of deposits price, the coefficient of labour price appears with a negative sign and is statistically significant. The results show that an increase in labour costs by one percent reduces interest revenue by about 0.17 percent. This may give an implication that banks are operating with more than an optimal amount of labour. Some studies have found a positive coefficient of labour price, Bikker and Haaf, (2002). The unit price of capital coefficient is negative and statistically significant. The unit cost of capital is the largest input contributing negatively to interest revenue. A one percent increase in operating expenses reduces interest revenue by about 0.56 percent.
Table 5.2 Regression Results of the Pazar-Rosse Model
VARIABLES ln(INTEREST REVENUE)
ln(Deposits Price) 0.618***
ln(Labour Price) -0.168*
ln(Capital Price) -0.561***
ln(Equity to total assets) 0.314***
ln(Loans to total assets) 0.733***
ln(Loan-Loss Provision) 0.220***
ln(Operating Income) -0.0143
Number of banks
Wald test: H=0 4
Notes: (***), (**) and (*) denote 1, 5 and 10 percent level of significance, respectively.
Standard errors are in parentheses.
The ratio of equity to total assets is positive and statistically significant. The result is in line with previous empirical studies such as (Simbanegavi et al, 2014 ; Mirza et al, 2016). The results also show that loans to total assets coefficient is positive and statistically significant at all levels of significance. This is in line with theory and expectations because an increase in bank loans implies higher interest revenue. The loan-loss provision appears with a positive sign and is statistically significant. Some studies have found a contradicting result for this coefficient. They argue that increases in non-performing loans reduce interest revenue. However, the reason for the positive sign of loan-loss provision may be that less risk averse banks generate more interest revenue as opposed to more risk averse banks. This is because more risk averse banks would grant loans only to customers with exceptional credit history, thereby making such banks earn relatively less revenue from loans and advances than less risk averse banks. Lastly, the results indicate that the ratio of other operating income to interest income appears with a negative sign, although it is statistically insignificant.
The H-statistic, which is the sum of coefficients of input prices, is -0.111. This statistic also represents the elasticity of banks’ revenues with respect to factor input prices. We perform Wald test to test if the H-statistic is statistically different from zero. The test fails to reject the null hypothesis at all levels of significance (p-value = 0.4106). Thus, we conclude that H-statistic is not statistically different from zero. This result indicates the case of a perfectly colluding oligopoly market structure. The H-statistic being not significantly different from zero implies a very low competition in the banking sector. This allows banks to have a high degree of market power. That is, they have ability to set uncompetitive prices and thus negatively affecting the consumers.
6CONCLUSIONS AND RECOMMENDATIONSThe thesis measures the degree of competition in Lesotho’s banking industry. To achieve this, the Panzar-Rosse (2012) methodology is used to find a measure of competition. The study uses bank level data spanning from 2013 to 2016 for the entire banking industry. The results suggest that banks in Lesotho operate under a perfectly collusive oligopoly, with the H-statistic of -0.1. The Wald test is was performed to check if the observed H-statistic is different from zero. The results suggest that the H-statistic is not statistically different from zero. This result is typical for a frim operating as a monopoly. The finding suggest that Lesotho commercial banks are acting collusively in setting their output and prices. The results of this study are different from other studies that measure competition in banking sector. Most studies found monopolistic competition, with moderate to high degree of competition as a prevailing market structure.
The findings of this thesis call for authorities to implement policies aimed at increasing competition in the banking sector. This is because low competition in the banking industry adversely affects consumers and hinders economic growth. It is important for authorities to encouraging participation of more players in the banking to reduce the level of concentration in the banking sector, as this is one of the concerns.
REFERENCESAktan, B. and Masood, O., 2010. The state of competition of the Turkish banking industry: An application of the Panzar-Rosse model. Journal of Business Economics and Management, 11(1), pp.131-145.
Bain, J.S., 1951. Relation of profit rate to industry concentration: American manufacturing, 1936–1940. The Quarterly Journal of Economics, 65(3), pp.293-324.
Barbosa, K., de Paula Rocha, B. and Salazar, F., 2015. Assessing competition in the banking industry: A multi-product approach. Journal of Banking ; Finance, 50, pp.340-362.
Baumol, W., Panzar, J. and Willig, R., 1982. Contestable markets and the theory of market structure. Nueva York, Harcourt Brace Javanovich, Inc.
Bikker, J.A. and Haaf, K., 2002. Competition, concentration and their relationship: An empirical analysis of the banking industry. Journal of Banking ; Finance, 26(11), pp.2191-2214.
Bresnahan, T.F., 1982. The oligopoly solution concept is identified. Economics Letters, 10(1-2), pp.87-92.
Bresnahan, T.F., 1989. Empirical studies of industries with market power. Handbook of industrial organization, 2, pp.1011-1057.
Claessens, S. and Laeven, L., 2004. What drives bank competition? Some international evidence. Journal of Money, Credit, and Banking, 36(3), pp.563-583.
Demsetz, H., 1973. Industry structure, market rivalry, and public policy. The Journal of Law and Economics, 16(1), pp.1-9.
Degryse, H., Moshe, K. and Ongena, S., 2009. Microeconometrics of Banking. New York: Oxford University Press
Fosu, S., 2013. Banking competition in Africa: Subregional comparative studies. Emerging Markets Review, 15, pp.233-254.
Freixas, X., Lóránth, G. and Morrison, A.D., 2007. Regulating financial conglomerates. Journal of Financial Intermediation, 16(4), pp.479-514.
Gujarati, Damodar. “Basic Econometrics Fourth (4th) Edition.” Magraw Hill Inc, New York (2004).
Hamza, R.A., 2011. Validation Panzar-Rosse Model in determining the structural characteristics of Tunisian banking industry. Journal of Economics and International Finance, 3(5), pp.259-268.
Karasulu, M., Fischer, R. and Meléndez, M., 2007. Competition in the Chilean Banking Sector: A Cross-Country Comparison with Comments. Economía, 7(2), pp.1-39.
Lau, L.J., 1982. On identifying the degree of competitiveness from industry price and output data. Economics Letters, 10(1-2), pp.93-99.
Mason, E.S., 1939. Price and production policies of large-scale enterprise. The American Economic Review, 29(1), pp.61-74.
Mathisen, Mr Johan, and Thierry D. Buchs. Competition and efficiency in banking: Behavioral evidence from Ghana. No. 5-17. International Monetary Fund, 2005.
Mirza, Faisal Mehmood, Olvar Bergland, and Isma Khatoon. “Measuring the degree of competition in Pakistan’s banking industry: an empirical analysis.” Applied Economics 48, no. 53 (2016): 5138-5151.
Okeahalam, C., 2002. Structure and conduct in the commercial banking sector of South Africa. Trade and Industrial Policy Strategies.
Panzar, J.C. and Rosse, J.N., 1987. Testing for” monopoly” equilibrium. The journal of industrial economics, pp.443-456.
Shaffer, S., 1982. A non-structural test for competition in financial markets. In Bank Structure and Competition, Conference Proceedings, Federal Reserve Bank of Chicago, 1982 (pp. 225-243).
Shaffer, Sherrill. “Competition in the US banking industry.” Economics letters 29, no. 4 (1989): 321-323.
Simatele, M., 2015. Market Structure and competition in the South African banking sector. Procedia Economics and Finance, 30, pp.825-835.
Simbanegavi, W., Greenberg, J.B. and Gwatidzo, T., 2014. Testing for competition in the South African banking sector. Journal of African Economies, 24(3), pp.303-324.
The World Bank. 2017. World Development Indicators. Washington, D.C.: http://data.worldbank.org/indicator/FS.AST.PRVT.GD.ZS?locations=LSAPPENDIX AFigure A1 Interest Rate Spread
Notes: Source: Authors’ calculations based on data from the World Bank data. Figure A1 depicts the interest rate spread in Lesotho’s banking sector for the period from 2000 to 2016. The lending spread is the difference between the lending and deposit rates.
Figure A2 Composition of Banks Revenue
Notes: Figure A2 depicts the aggregate industry composition of total revenue. The amounts are in millions of maloti.