Impacts of Bank Competition, Financial Stability, and Gender Gap on Access to Finance: Evidence From Sub-Saharan Africa

Using a dataset of 12504 rms from World Bank Enterprise Survey (WBES), this paper investigates the role of bank competition, nancial stability and gender gap in access to nance in Sub-Saharan Africa. We empirically test the existence of market power hypothesis according to which pro-competitive policies alleviate credit constraints from the banking industry. Results obtained through probit model and probit model of sample selection (PSS) conrm that a higher degree of market power negatively affected rm nancing in the region. Findings further reveal that the lower rate of female ownership partnerships creates diculties in obtaining formal nance for female entrepreneurs. Financial stability does not affect access to nance in SSA as indicated by the estimation results. We discuss several policy implications for the region.


Introduction
Limited access to nance is considered a major challenge to economic growth both in developed and emerging economies (Demirguc-Kunt et al. 2008). It is argued that a deep banking system can limit access to credit if it lacks competition, serves only incumbent rms and operates without regard to prudential regulations. While extant literature documents the implications of bank competition in promoting economic growth (Aluko and Ajayi 2018;Cherif et al. 2020;Sarpong-Kumankoma et al. 2020), adequate attention has not been paid on the implications of bank competition and nancial stability in rm's nancing constraint. Financial liberalizations in the 1990s, the boom in commodity prices, and economic growth have contributed signi cantly to the expanding banking sector in Sub-Saharan Africa (SSA) (Mlachila et al. 2013;Otchere et al. 2017). Despite the realized merits of bank competition, the structure of markets is often characterized by anti-competitive practices and regulations. The widespread presence of state-owned banks, the greater degree of monopoly and holding the largest market share by single operators are the key elements that affected the banking competitiveness in the region. Even though a decrease in competition and rising corporate market power has gained notable attention in emerging economies (IMF 2019a;De Loecker et al. 2020), a systematic analysis in the pattern and variation of bank competition in SSA remained unexplored. Based on the sample of selected SSA countries, Cherif et al. (2020) show that in general, bank competition in SSA remains low compared to the rest of the world.
In addition to low bank competition, SSA region is also grappled with the problem of nancial stability. In the face of the Nigerian banking crisis of 2009-10, increased protectionism, stricter regulations in global nancial conditions, and external pressures including the threat of de-risking have raised concern over nancial stability in the region. Previous studies mostly focus on the effects of nancial inclusion on nancial stability, ignoring the likely possibility of the role of stability in channelizing access to nance to rms. Considering the fact that a competitive banking system may not necessarily stable (Berger et al. 2017), this study attempts to examine the joint impact of competition and stability on rms' credit availability.
Another less debated aspect in the discussion of access to nance is the potential gender gap. It is argued that lack of access to nance to female-led business enterprises impedes female entrepreneurship and discourages female participation in the modern nancial market. With less than one in ve households having access to nance, this problem is much stressing in SSA than in other developing countries (Honohan and Beck 2007). World Bank data shows the SSA region is characterized by a higher degree of gender discrimination measured by Women, Business and the Law index as compared to other countries in the world. The values of the index are substantially lower in SSA that the rest regions. It indicates that the average index value in SSA is 0.5 and 0.85 for the rest countries. Smaller values indicate higher gender discrimination; it can be inferred that the region's business and legal environment is characterized by a higher degree of the gender gap.
Given the background, this study seeks to address the following research questions. First, does an increase in bank competition enhance or hindrance access to formal nance in SSA region? Second, whether ensuring nancial stability improves rm's nancing. What is the joint effect of bank competition and nancial stability on credit availability? Third, does gender discrimination affect the probability of credit constraint for female-led enterprises?
This study contributes to the existing literature on rm nancing in the following ways. Using rm-level data from Word Bank Enterprises Survey (WBES) and bank-level information, this study investigates the role of competition, nancial stability, and the gender gap in selected SSA. While most studies on access to nance are limited to developed and emerging economies, this study is the rst of its kind that explores the association between market power and credit availability in SSA. Aterido et al. (2013) considered the role of the gender gap in analyzing access to nance, however, the implication of bank competition and nancial was not addressed in their study. Since existing studies revolve around two competing hypotheses viz., market power hypothesis and information hypothesis to examine the implication of bank competition on access to nance, this study tests the existence of these two hypotheses in the context of SSA region. Based on the income level, we further analyzed the variations in rm's credit status across the countries in the region. Finally, we apply both direct and indirect measures of access to nance, several rm-speci c, country-speci c, macroeconomic, and institutional variables in this study.

Methodology And Data
First, we measure bank competition and nancial stability by Lerner index, Boone indicator and Z-score. Lerner index is formally de ned as the relative difference between price and marginal cost divided by the price and corresponds to the inverse of the price elasticity of demand. Lerner index (1934) can be measured as follows.
where P i,t denotes the price of bank i at time t. MC denotes the marginal cost of banks and e is the elasticity of demand. The value of the Lerner index lies from a maximum value of one to the minimum value of zero. Boone (2008) de nes market power as an estimate of the percentage decrease in pro ts resulting from one per cent increase in marginal cost. Boone indicator for bank i at time t can be de ned The inverse relationship between pro t and marginal cost suggests the pro t elasticity to be negative.
The stronger the effect, the more negative value of Boone indicator is.
Z-score measures the probability of default of the banking system. Following Cihak and Hesse (2010), we estimate the Z-score as Where ROA denotes return on assets calculated by bank's net income divided by total assets. E/TA is the equity to total assets, and σ ROA is the standard deviation of the return on assets. A higher index of Zscore implies that banks are relatively more stable and have a lower probability of being insolvent and vice versa. We use three years rolling window to estimate the standard deviation.
After estimating bank competition and stability indicator, following Aterido et al. (2013), we specify the following regression equation, where our interest of variables are measured at the level of rm i in country c at time t The dependent variable is a dummy variable y that takes a value of 1 if a rm has access to a bank loan or overdraft facility, or 0 otherwise. To capture the role of gender, we rst consider the participation of female ownership and assign a dummy variable 1, indicating at least one principal owner is female. We also consider the gender of the manager and assign a dummy variable one if the manager is female. FS that denotes rm-speci c characteristics includes the size of the rms by the number of permanent fulltime employees. We consider three types of rms, namely small rms (6 to 10 employees), medium rms (11 to 49 employees) and large rms (50 or more employees). Experience is de ned by the experience of the top managers in years. Additionally, we consider the age, ownerships and sector of the rms (manufacturing and service rms only). To get rid of the omitted variable bias, we take into account a set of control variables such as bank branch, nancial freedom and macroeconomic variables, namely GDP growth and in ation rate. Survey xed effects are denoted by M. Survey xed effects are crucial to control for the potential shocks or measurement errors across different country years, where the surveys are conducted. Our sample includes 40 SSA countries covering 12504 formal rms during the period 2009 to 2019 (Table 2a). We estimate equation (4) with probit model and PSS. Table 1 presents the descriptive statistics of the variables. It shows that 67.40 per cent of total rms in our sample have access to nance. Considering the alternative indicator, we report that 84.5 per cent of rms face credit constraints. This implies that on average, 84.4 per cent of total rms despite having a formal credit need did not apply for a formal bank credit due to various reasons. The maximum value of the Lerner index (49.5) indicates the existence of higher market power and less bank competition in the region. With regards to the gender of the owners, we nd that only 12.1 per cent managers are female, and 23.1 per cent of rms have at least one female owner. The mean value of the size variable suggests that most rms are medium-sized.  2 show that Lerner index has a negative and signi cant impact on access to nance. One per cent increase in market power causes access to nance to fall by 3.7, 2.9, 4.4, and 4.1 per cent respectively across all the estimated models. Higher interest rates charged on loans due to the banking industry's growing monopoly signi cantly reduce credit availability. In other words, bank competition has the potential to improve the access the nance by considering the low-interest policy. The nding lends support to market power hypothesis and is in line with Leon (2015) and Ayalew and Xianzhi (2019). Though we do not nd any signi cant impact of nancial stability on access to nance, the coe cient of the interaction term between the Lerner index and Z-score has been found positive and signi cant across all the models. This allows us to infer that a stable and competitive banking industry helps improve access to nance in the region Less participation of females in business enterprises makes it di cult for female-led entrepreneurs to receive formal credit. This nding has been con rmed by the negative and signi cant coe cient of female ownership in all the estimated models. However, we fail to draw any inference about the role of the female manager in obtaining rm nance. Though the coe cient sign is negative as expected, lack of signi cance level restricts us to derive de nitive implications of whether female managers face di culties in obtaining nance. The ndings further reveal that small and medium-sized rms face greater credit constraints compared to their large counterparts. Privately-owned rms have greater credit constraints than government rms. The signi cant coe cients of bank branches suggest that the widespread presence of bank branches per 100,000 adults is likely to increase access to nance. With regards to the macroeconomic and institutional factors, we nd consistent results in line with the economic intuitions.      To see the extent to which our main ndings are consistent, we check the robustness of the main estimation results. To do so, we replace the dependent variable with an alternative indicator, i.e. credit constraint. This measure is only relevant only for the rms with a need for bank credit. Therefore, we control for the possibility of sample selection bias by applying the probit model with sample selection (PSS). While performing the robustness test, we use the Boone indicator instead of the Lerner index as a measure of bank competition. Table 3 presents the results of the robustness. Table 3 shows that Boone indicator, a direct measure of market power exerts a positive and signi cant impact on credit constraint, implying that market power gives rise to credit constraint in the SSA region. Similarly, we nd that female ownership participation has a positive and signi cant impact on credit constraint. The economic implications of the negative ndings can be attributed to the weak credit information system in SSA. Lack of credit information may create informational asymmetries in the credit market, leading to adverse selection and moral hazard problems, which makes external nance less affordable to the borrowers.

Empirical Results
Finally, we examine the impact of bank competition, nancial stability and gender role on access to nance among the low income, lower middle income and upper-middle-income countries. The results have been presented in table 4. We group our total sample countries into the three income groups. We nd that higher market power invariably affects rms nancing across three groups. However, unlike upper middle income, female ownership participation adversely affects access to nance in lower and lower middle income countries of SSA. Lastly, we nd that small and medium-sized rms face greater nancing obstacles in lower and lower middle income countries.

Conclusion And Policy Implications
This study concludes that a higher degree of market power in the SSA region has adversely affected access to nance, supporting the market power hypothesis. Findings do not extend any support about the effect of nancial stability on rm nancing. However, the joint impact of competition and stability suggests that a higher degree of competition and nancially stable banking industry can promote formal credit access. The gender gap appears to have played a crucial role in obtaining nance as fewer female participation, and female managers nd it di cult to raise credit from the banking markets.
The ndings of the study suggest the following policy implications. First, the banking industry in the SSA region, in general, should pursue pro-competitive banking policies. Higher bank competition by lowering interest rate margins can signi cantly improve access to nance. Second, government stakeholders should pay due attention to reduce the gender gap from every sphere of the region and ensure greater participation of women in business enterprises. Financial policies should be designed in such a way that it encourages women in the model nancial markets. Third, policymakers need to ensure that competitive banking operates in a stable nancial system.

Declarations
Availability of data and materials We use the data from World Bank Enterprise Survey (https://www.enterprisesurveys.org) and World Bank, Global Financial Database (https://www.worldbank.org/en/publication/gfdr/data/global-nancialdevelopment-database). The dataset used in this study is available upon request from the author.