Lending Channel of Monetary Policy: Does Market Power Matter? Evidence from South Asia

1.     Introduction

The banking industries of South Asian economies have exhibited significant changes over the last few decades. While the speed and extent of these reforms have varied from country to country, financial liberalisation fosters competition among financial institutions. Financial market openness has attracted new investments through privately owned banks and has encouraged many foreign banks to set up branches in South Asian countries. These financial developments have resulted in a sharp increase in competition among banks while reducing the market share (market concentration) of the state-owned banks. On the other hand, the global financial crisis in 2008 caused the financial regulators of most countries to encourage small financial institutions to consolidate with each other, as a part of the restructuring process that aimed to increase the soundness of the banking sector and thus stabilise the financial system. Similarly to the contradictory views on competition-fragility and competition-stability, the literature presents opposing findings on the impacts of market power and competition on monetary policy transmission in any economy.^[1]The global financial crisis reignited the interest of policymakers and academics in bank competition and the role of the state in developing competition policies (World Bank 2020b). This confirms the necessity of establishing the correct laws and policies that determine adequate bank competition in the financial market, to ensure social welfare.

The bank lending channel of monetary transmission operates via the supply side of the credit market, rather than the traditional demand-side of monetary transmission, namely the interest rate channel (Olivero et al. 2011b, p. 1035). Changes in banking market power, and in particular concentration/competition in the banking market, are expected to impact the transmission of monetary policy, primarily through the lending channel.  The market power no doubt alters the supply of loans in the market directly. However, it may also impact the loan supply indirectly via the monetary transmission mechanism. Some studies have already examined the relationship between the bank market concentration/competition and the lending channel. However, the existing literature does not provide conclusive evidence in the context of developed and some developing countries.

With the exception of a few studies of the effects of bank competition on bank efficiency, revenue diversification and pricing, no empirical studies on market power or market concentration in the South Asian banking sector seem to be available. Apart from this gap in the existing literature, there are several reasons for this essay.  First, this is one of the few cross-country studies to focus on the banking industry and its impact on monetary policy in the emerging market economies of South Asia. Banks play an essential role in these countries’ financial systems, as most economic agents depend on bank lending as a primary source of finance. This is due mainly to the underdeveloped capital market and the lack of availability of active secondary markets. Second, the majority of the previous studies on monetary policy transmission channels have relied on aggregate macro-level data, which raises the concern of the identification problem.^[2] This paper uses bank-level data to examine the effects of bank concentration on supply-side monetary transmission from the lending channel, as opposed to the demand-side interest rate channel of monetary policy transmission in South Asian economies. Moreover, the use of bank-level data allows the systematic differences in the effects of bank concentration across different balance sheet strengths and corporate attributes to be examined. Third, in the aftermath of the global financial crisis, bank regulators encouraged mergers and acquisitions among financial institutions as a prudential policy in order to maintain financial stability.  As has been documented in the literature, a higher concentration/lower competition in the banking industry may have either a positive or a negative impact on monetary policy transmission. Hence, the findings of this essay can be used to assess the use of bank market concentration as a stability measure, relative to its impact on the effectiveness of monetary policy.

For these reasons, this paper aims to revisit the lending channel of monetary policy transmission in South Asian economies and to examine the effects of bank concentration on monetary policy. The study uses bank-level panel data from five South Asian economies, namely Bangladesh, India, Nepal, Pakistan, and Sri Lanka, over the period 2011–2018. Specifically, this paper addresses two key questions: (i) Does the bank lending channel of monetary policy transmission exist in these economies? and (ii) Does banking concentration have an impact on the bank lending channel of monetary policy transmission?

The study employs five non-structural measures for inferring the level of bank concentration: three HHIs, defined as the market share of the banking sector in terms of total assets (HHI-TA), total loans (HHI-L) and the deposit base (HHI-D), and the three-firm and five-firm concentration ratios (CR3 and CR5 respectively).  Using the fixed effects (FE) model, the analysis provides new evidence on the presence of the bank lending channel of monetary policy transmission. The results suggest that monetary policy tightening induces banks in South Asian economies to reduce their loan supply. Furthermore, they suggest that a higher concentration in the banking industry serves as a buffer in the transmission of monetary policy shocks to bank lending, meaning that monetary policy transmission becomes weaker and less effective.  Moreover, the weakened effect is more evident for smaller, less-liquid, poorly-capitalised and less-profitable banks. This is consistent with the existing literature on the bank lending channel, which shows that banks which are more financially constrained find it difficult to substitute loanable funds in a monetary contraction. These findings are broadly robust across alternative measures of monetary policy and the different estimated methods. From a policy standpoint, our findings call for a close oversight of banking concentration and the development of measures that can offset the negative effect of banking sector concentration on the effectiveness of the monetary policy transmission mechanism.

The rest of this paper is organised as follows. Part 2 of this paper provides a brief literature analysis of the bank lending channel and the theory as to the way in which market power impacts monetary transmission. Part 3 discusses the empirical strategy and describes the data. Part 4 begins by estimating the baseline model and then proceeds to estimate an alternative model for ensuring the robustness of the baseline results, and part 5 provides the conclusions and policy implications of this paper.

2.     Literature Review

Following the traditional IS-LM model, a decrease in the money supply has an immediate impact on the money market, increasing the equilibrium interest rate in the money market. It then impacts the real decisions of firms and households through a reduction in investments. In line with this process, a decrease in deposits on the liability side of a bank’s balance sheet is set off by a proportionate reduction in loan disbursements and bond investments on the assets side of the balance sheet. This is called the ‘money view’ of monetary policy  (Taylor 1995). Monetary policy changes assets’ prices and returns, and thus influences the bond price, interest rates and exchange rates. Under this view, money plays a unique role, but banks have no active involvement except in the issuing of deposits, as bonds and loans are considered as perfect substitutes (Lensink & Sterken 2002).

In contrast, the ‘credit view’ considers credit and bonds as imperfect substitutes. In this case, some banks will prefer to build a bond portfolio and slow down their loan disbursements following a change in monetary policy.^[3] Hence, banks play an active role in monetary transmission as providers of credit to the economy (Bernanke & Gertler 1995; Cecchetti 1995). The existing literature on the credit view has two sub-channels of such transmission: the bank lending channel and the balance sheet channel.  Since the influential works of Bernanke & Gertler (1995) and(Kashyap & Stein 1995), several studies have confirmed the existence of the lending channel of transmission. Those studies explain how monetary policy changes affect bank balance sheets through the variation in the supply of loans (See  Altunbaş et al. 2002; Brissimis & Delis 2009; Kashyap & Stein 2000; Kishan & Opiela 2000). In general, the underlying assumption with bank lending is that monetary tightening drains banks’ reserves and deposits, and this reduction in loanable funds causes banks to shrink their available loan portfolios. Similarly, an expansionary monetary policy replenishes banks’ reserves and deposits and leads to a spike in their loanable funds.

However, if banks can replace the reduction in loanable funds using an alternative source of funds (e.g. the issuance of new equity) in the case of monetary tightening, the lending channel will break down (See Fungáčová et al. 2014). As most of the recent literature points out, this is not a reasonable assumption due to various industry-specific features of the banking business. Banks’ responses to monetary policy changes depend on their balance sheet characteristics (Kashyap & Stein 2000). (Kishan & Opiela 2000) suggested that the size of a bank’s assets is associated significantly with the bank’s reaction to the monetary policy shocks. They argued that monetary tightening has a stronger impact on small banks due to their lower ability to access alternative sources of funding. Peek & Rosengren (1995) found that bank leverage, represented by the capital to assets ratio, is an essential factor for determining the sensitivity of bank loans to monetary policy initiatives. Similarly, the literature indicates that well-capitalised and highly-liquid banks are more resistant to monetary contractions, as they have access to unsecured funding in the market in order to protect their loanable funding (Juurikkala et al. 2011; Kishan & Opiela 2006). Hence, as was discussed earlier, banks will be reluctant to expand their lending if they find it costly or difficult to raise capital, even if there is a higher demand for loans in the market.

Recently, the modelling of banks’ reactions to monetary policy impulses has been modified slightly by Disyatat (2011) through the external finance premium hypothesis. The tightening of monetary policy leads to a deterioration in asset quality and banks’ leverage, which increases the external finance premium. Changes in this risk perception of a bank’s balance sheet increase the cost of funding due to a higher external finance premium. In turn, banks pass this premium on to customers by increasing the lending rate (Cantero-Saiz et al. 2014). From an external finance premium perspective, the mechanism of the bank lending channel is that a contractionary monetary policy reduces a bank’s loanable funds through a fall in deposits and/or an increased external finance premium, thus limiting the loan supply to borrowers.

While there has been ample research on the lending channel, the majority of the studies have focused on developed countries, and the results on the existence and effectiveness of the lending channel have been inconclusive. In general, the results suggest that small banks with poor capitalisation and weak liquidity positions are more likely to have their loan supply affected adversely by monetary policy contractions, as they find it difficult to obtain alternative sources of finance. Hence, banks with weak balance sheets opt to downsize their lending portfolios. Unlike developed economies, research results on developing countries provide more conclusive evidence on the existence of the lending channel (See Freedman & Click 2006; Sanfilippo-Azofra et al. 2018). This is because banks are the primary financial intermediary in developing economies, and deposits are their primary source of funding for the provision of credit facilities. Hence,  monetary tightening has a negative impact on the loan supply (See De Mello & Pisu 2010; Hou & Wang 2013). These studies further suggest that, while the lending channel is present in developing economies, its intensity in developed economies varies.

3.     Empirical Strategy

3.1. Data

This study uses annual bank-level data for five selected South Asian economies: Bangladesh, India, Nepal, Pakistan, and Sri Lanka. The panel data encompass the unconsolidated balance sheets and income statements of 125 commercial banks, including 906 bank-year observations.[4] The sample period is from 2015 to 2022, covering the systemic important commercial banks of selected countries.  The data are obtained from several sources, with unconsolidated bank-level financial statements being extracted from the Orbis database, provided by Bureau van Dijk.[5]

The sample includes only domestic-commercial banks, to ensure the comparability of the data. All other financial institutions, such as specialised banks, investment banks and finance companies who engage in similar types of business, are excluded, as the regulatory provisions and reporting requirements applied to these institutions may differ from those of commercial banks. Foreign-owned commercial banks are excluded due to the lack of availability of unconsolidated financial statements. In cases where corporate mergers or acquisitions took place during the sample period, the acquiree/target bank has been eliminated from the sample, and the acquirer bank is treated separately by considering unconsolidated data in the sample. Similarly, banks with short periods of financial statements and banks with outliers are also eliminated from the sample.^[6] The country-specific macroeconomic data are extracted from the financial statistics published by the International Monetary Fund (IMF) and the world development indicators published by the World Bank. This section presents the descriptive statistics and correlations among the main variables used in the study, a summary of which is provided in Table A1 in the appendix.

3.2. Variables in the Model

This study employs five alternative measures of the concentration ratio: three HHIs, defined as the market share of the banking sector in terms of total assets (HHI-TA), total loans (HHI-L) and the deposit base (HHI-D); and the three-firm (CR3) and five-firm (CR5) concentration ratios. As was noted earlier, the market share of large banks, measured by the CR3 and CR5 of these five economies, has decreased over the period under investigation.

Changes in bank lending activities are represented by the annual percentage change in the bank’s gross loan portfolio. This is the dependent variable that is used most commonly in the literature for studying the lending channel of monetary transmission.  The size of the domestic commercial banking market varies across the selected economies. Hence, this study avoids the size effect of loans in larger markets by using the growth of loans instead of the volume of loans, as was suggested by Olivero et al. (2011b).  Following the monetary policy literature, I measure changes in the monetary policy stance by changes in the short-term interest rate; i.e., a monetary policy easing (tightening) is reflected in a drop (hike) in the short-term interest rate (See Bernanke & Blinder 1988). I mainly use changes in the money market rate (MMR) to represent the monetary changes in this study.

Most of the early empirical studies on the bank lending channel have since been challenged, as they relied on aggregate data. The issue with using aggregate data for the empirical estimates is that it creates an ‘identification problem’, namely a difficulty in identifying the exact reason(s) for the reduction in the loan supply. In other words, using aggregate data to study changes in loans cannot separate the demand-side and supply-side effects of monetary policy transmission.  Kashyap & Stein (1995) were the first to use disaggregate data to address the identification problem. Since then, many researchers have pointed out that the decline in bank loans following a monetary policy tightening is tied closely to individual bank characteristics, such as liquidity position and capital strength (Gambacorta 2005; Kashyap & Stein 2000). The other side of this argument is that all banks face identical loan demands, which suggests that the demand for loans does not depend on bank characteristics (Fungáčová et al. 2014). For example, following the introduction of an expansionary monetary policy, the loan demand decreases by the same amount for both small and big banks (see Bernanke & Blinder 1988). While recent empirical studies using loan-level data to study the effect of monetary transmission have relaxed this assumption (Jiménez et al. 2012), our study retains this assumption because loan-level data are not available in the South Asian context. Furthermore, most South Asian countries still rely on the bank-based financial system; customers do not have an alternative to bank loans as  a source of finance.

The bank-level data used in this analysis allow me to control for the degree and type of financial strength or constraints of heterogeneous banks (Ashcraft 2006). Bank size, capitalisation, and liquidity are vital bank-specific characteristics that are associated with information asymmetries (Kashyap & Stein 2000). Hence, I include several bank balance sheet variables, to control for the effects of bank balance sheet conditions on changes in credit growth and supply that are unrelated to the monetary policy changes. I measure the individual bank size by the logarithm of total assets, as bigger banks could easily handle adverse monetary shocks by issuing financial instruments. The liquidity position of each bank is computed as its ratio of liquid assets (cash, deposits with other banks and short-term securities) to total assets, and its degree of capitalisation is computed as the equity to total assets ratio. These three prominent bank characteristics influence the accessibility of and the premium on external finance. Banks with better capital and highly liquid assets tend to pay lower risk premiums for their debt, and can still create loans against unexpected deposit shocks caused by monetary policy changes.

I also include GDP growth and financial sector development, to control for macroeconomic changes and the availability of alternative sources of finance, respectively. These two factors may affect the demand for bank credit from the banking customers in the selected economies. Controlling for demand-side effects helps to distinguish the supply-side effects of the lending channel. Furthermore, no significant financial or monetary sector reforms occurred during the sample period.  Table A2 in the appendix presents the definitions and the data sources in more detail.

3.3. Models and Estimation Techniques

I uncover the effect of the banking market concentration on the lending channel of monetary policy transmission by following the empirical approach introduced by Kashyap & Stein (2000). This approach has been used subsequently to analyse the lending channel in both developed and emerging market economies (Bhaumik et al. 2011; Gambacorta 2003; Khan et al. 2016). The specification of the model is

where the subscripts i, c and t denote the individual bank, the country in which the bank operates and the time, respectively. The outcome variable   represents the growth rate of bank loans for bank i, in country c, at time t.  is the lagged value of the outcome variable.  is a measure of the change in monetary policy.  is the vector of control variables for bank-specific characteristics, including size, capitalisation, deposit growth, profitability, liquidity, risk of problem loans, and capital adequacy.  represents country-specific macroeconomic characteristics, including GDP growth and financial sector development.  is the bank fixed effects,  is the year fixed effects and  is the error term.

I model the effect of a bank’s market structure on the lending channel of monetary transmission by extending Eq. (3.1) to include the market concentration term and the interaction term of the bank concentration with the monetary policy measures, where the extended model is given by

Here,  measures the bank market concentration for each country-year. The interaction term  captures the marginal impact of the banking sector market concentration on monetary policy transmission through the bank lending channel. The remaining variables are the same as in Eq. (3.1). Guided by prior studies on the bank lending channel, all control variables are lagged by one year to avoid an endogeneity bias (See Kashyap & Stein 2000; Kishan & Opiela 2000).

The coefficient on the bank concentration ( captures the effect of the market concentration on the loan growth. Similarly, the coefficient on monetary policy  estimates the response of the loan growth of bank i to monetary policy impulses.  As per the literature, it is expected that an increase in the monetary policy rate will lead to a reduction in bank lending. Hence, the  coefficient, which represents the effect of monetary policy on the lending growth, should be negative. As was discussed in the literature review section, banking sector concentration can either weaken or strengthen monetary policy transmission via its impact on bank lending. Therefore, the coefficient of the interaction term (), which shows the marginal effect of market structure on the banking lending channel, can be either negative or positive. A positive (negative) coefficient indicates that the sensitivity of bank lending to monetary policy is smaller (larger) when the concentration in the banking industry is high. This suggests that the market structure weakens (strengthens) monetary policy transmission via the bank lending channel.

The empirical method used in this paper is based on previous studies of bank lending channels. The majority of those studies follow the empirical framework developed by Kashyap & Stein (2000), which examines whether individual banks react differently to monetary policy changes (for instance, Bhaumik et al. 2011; Fungáčová et al. 2014; Hou & Wang 2013; Olivero et al. 2011b). I use the fixed effects estimator as my main estimation method, then confirm the robustness of the empirical results using an alternative estimation method. Specifically, I re-estimate the model using a System Generalised Method of Moments (GMM) method. GMM is a popular method of dynamic panel estimation in the banking and finance literature because it ensures more efficient and consistent results. I then split the data into different subsamples based on different bank-specific characteristics in order to examine the relationship between bank concentration and monetary policy transmission across heterogeneous banks.

4. Empirical Results

The regression results for Eq. (2), estimated using the fixed effects model, are reported in Table 1. The dependent variable is the growth rate of loans in all specifications. I use five different measures of market concentration to capture the effects of the market structure on monetary policy transmission. Panel A presents the results of Eq. (2) without the interaction term, while panel B includes the interaction term.

As expected, I find the coefficient of monetary policy to be negative and significant in all specifications. The negative coefficient suggests that an expansionary (contractionary) monetary policy, characterised by a downward (upward) adjustment to the interbank money market interest rate, has a positive (negative) effect on the credit supply in all economies. Hence, monetary policy contractions induce banks to cut down their credit supply and suggest the existence of a bank lending channel in South Asian economies. This result is consistent with prior empirical evidence on the lending channel of monetary policy in South Asian countries, such as Bhaumik et al. (2011) and Bhatt & Kishor (2013) for India, Afrin (2017)for Bangladesh, and Perera et al. (2014) for South Asian economies in general.

The coefficients on the market concentration measures are significantly negative except in model 9, which suggests that the supply of loans grows at a slower rate in more concentrated markets. Furthermore, the results in Panel B of Table 1 show that the coefficient on the interaction term between bank concentration and monetary policy is positive and statistically significant except for CR3 and CR5. This result indicates that banks with higher concentration/market power are less sensitive to monetary policy changes. Thus, an increased concentration/increased market power weakens the transmission of monetary policy through the bank lending channel. This may be due to higher market power enhancing the accessibility of the interbank money market and/or an alternative source of funds.

On the other hand, it is more difficult and costly for banks with low market power to access liquidity and funds. Such banks are more vulnerable to monetary policy shocks and have limited ability to hedge their lending activities against monetary policy shocks. Higher competition in the banking sector ensures that changes in monetary policy influence the cost and availability of funds. Hence, an increase in bank competition may facilitate a more direct transmission of monetary policy to the supply of bank loans. These results are consistent with the results of Olivero et al. (2011b) for the Asian and Latin American contexts.

Table 1: FE Estimates with and Without the Interaction Term

Note: This table reports panel estimates with bank and time fixed effects. The dependent variable is the loan growth rate, and the monetary policy variable is the change in the money market rate. All control variables are lagged one period. Standard errors clustered by bank ID and year are in brackets. , * and *** denote significance at the 10%, 5% and 1% levels, respectively.

Using the results reported in Table 1, I next compute the overall impact of a one per cent change in monetary policy on bank lending at different levels of market concentration. Table 2 reports the changes in bank lending, obtained by . Columns 1, 3, 5, 7 and 9 in the table show the level of concentration at different percentiles, while columns 2, 4, 6, 8 and 10 report the percentage changes in bank lending in response to a one per cent change in monetary policy for different concentration measures. The table shows the role of the market structure at different intensities for alternative concentration measures. These results show that an increase in the interest rate has a negative impact on lending, but that a higher concentration in the banking sector reduces the marginal effects of monetary policy.

Table 2: Percentage Change in Bank Lending Following a One Per Cent Increase in Monetary Policy

Note: Values are calculated as , using the results obtained in Table 1.

When an economy has a concentration at the 10^th percentile of the HHI-TA, a one per cent increase in the MMR induces a 0.0206% reduction in the supply of loans, while for HHI-TA at the 50^th percentile, the reduction in the supply of loans decreases to 0.0201%. Similarly, for concentration measured at the 75^th percentile, the decrease in the loan supply in response to a one per cent monetary policy shock drops to only 0.0128%. Similar analyses are also conducted for all other measures of the market concentration, but the results are consistent across alternative concentration measures and therefore are not reported here for the sake of brevity. These results show that an increase in monetary policy (tightening) always has a negative impact on bank lending, and decreases the supply of loans. However, the real effect of monetary policy diminishes as the banking sector grows more concentrated. A higher concentration ratio also decreases the loan growth, a result similar to that found in the empirical analysis conducted by  Khan et al. (2016) for the ASEAN economies.

In summary, the regression results show that a higher concentration in the banking industry serves as a buffer in the transmission of monetary policy shocks to bank lending; thus, the transmission of monetary policy becomes weaker and less effective as the banking industry grows more concentrated. 

4.1. Robustness Tests

Robustness With Time-Varying Fixed Effects: The first robustness check involves the addition of time-varying fixed effects to the model. Accordingly, I incorporate a country-year interaction term into the model in addition to the bank and time fixed effects. As Table 3 shows, the coefficient on monetary policy is negative and statistically significant. However, it is much higher in magnitude than the corresponding value in the baseline model. Under this specification, the coefficient on market concentration is positive and significant, suggesting that a higher market power enables banks to increase their loans. The interaction term between the monetary policy indicator and the market structure is still negative and significant, although with a different magnitude. Similarly to the earlier findings, the effect of monetary policy transmission is dampened with a higher rate of bank concentration. Overall, these results seem to be consistent with those in Table 1 using a panel FE without controlling for the time-varying heterogeneity.

Table 3: Robustness Check: Model With Time-Varying Fixed Effects

Note: This table reports panel estimates with bank and time fixed effects. The dependent variable is the loan growth rate, and the monetary policy variable is the change in the money market rate. All control variables are lagged one period. Standard errors clustered by bank ID and year are in brackets. , * and *** denote significance at the 10%, 5% and 1% levels, respectively.

4.2. Robustness: Alternative Measures of Monetary Policy

I test the robustness of the estimated results further by using an alternative measure of the short-term interest rate. The literature uses several different interest rates to measure monetary policy changes. I use the three-month Treasury bill rate (TBR) as a replacement for the money market rate, which is used in the baseline model as the indicator of monetary policy, and Table 4 displays the results. The three-month TBR is a standard proxy for monetary policy in the literature on the bank lending channel.

The results are similar to those obtained from the baseline model. First, the coefficient on monetary policy is significantly negative for all concentration measures. The growth in loan supply decreases in response to a monetary tightening, as it leads to increases in both of these interest rates. Hence, the existence of the lending channel of monetary policy transmission in these economies is still supported. Second, the interaction term is positive in all models and significant in all except the HHI-TA and CR3 models.

Table 4: Robustness Check Using Alternative Monetary Policy Measure (Treasury Bill Rate)

Note: This table reports panel estimates with bank and time fixed effects. The dependent variable is the loan growth rate, and the monetary policy variable is the change in the short-term Treasury bill rate. All control variables are lagged one period. Standard errors clustered by bank ID and year are in brackets. , * and *** denote significance at the 10%, 5% and 1% levels, respectively.

4.3. Robustness: Alternative Estimation Methods

To further validate these findings, I use an alternative estimation method to check the robustness of the empirical results obtained above. In the bank lending literature, some scholars prefer to estimate the dynamic model in Eq. (2) using the system GMM introduced by Arellano & Bond (1991). Accordingly, I re-estimate the model specified in Eq. (2) using the system GMM with robust standard errors. Our main findings stay intact.

The results of the FE model show that the lagged value of the loan growth is not significant. This may be due to the fact that annual rather than quarterly data are employed for the analysis. However, only annual data are available in the South Asian banking context. Furthermore, scholars criticise the economic rationale for using lagged loan growth in the model, as the current lending growth does not influence next year’s lending growth directly. They argue that the lagged value of the loan growth is relevant at monthly and/or quarterly data intervals (See Fungáčová et al. 2016; Fungáčová et al. 2014).^[1]

The system GMM estimates have a lower bias and higher efficiency for small samples. Further, endogeneity problems may arise from certain bank variables (like size, liquidity, and capital). Hence, the system GMM is intended to check the consistency of the results obtained using the baseline model on the issue of endogeneity. In the system GMM, the first equation is the first- difference equation, and the second equation is the level equation. When estimating the system GMM, the monetary policy indicator and the macroeconomic variables are considered to be exogenous, while the bank characteristics are endogenous. The instruments used are the first and second lags of the dependent variable and the first lag of all explanatory variables. I validate the model by conducting the autocorrelation and Sargan/Hansen tests in order to explore the issues of serial correlation and the validity of instruments, respectively.

Table 5: Robustness Check: Model Re-estimated Using the System Generalised Method of Moments (GMM) Method

Note: This table reports the GMM estimate. The dependent variable is the loan growth rate, and the monetary policy variable is the difference in the money market rate. All control variables are lagged one period. Robust standard errors are in brackets. , *, *** denote significance at 10%, 5% and 1% respectively.

The results of the system GMM method are presented in Table 5.  Our regressions pass the Arellano-Bond test for second-order serial correlation and the Sagan test for the overidentification of restrictions. Consistent with the FE model, it finds that the lagged value of loan growth is not significant.  These results are also broadly consistent with those from the baseline model.

Overall, this study has obtained several robust findings thus far.  First, the negative and statistically significant coefficient on monetary policy confirms the existence of the bank lending channel, emphasising the fact that monetary policy tightening induces banks in the South Asian economies to reduce their loan supply effectively. Second, the negative and significant coefficients on the concentration measures suggest that the loan supply grows at a lower rate in more concentrated markets. Finally, the positive and significant coefficient on the monetary policy and concentration interaction term suggests that an increased concentration weakens the monetary policy transmission and decreases the loan supply.

4.4. The Heterogeneity Effects

It is agreed widely that banks with different degrees of financial constraints respond differently to monetary policy shocks (Kashyap & Stein 1995; Kashyap & Stein 2000).  Therefore, I examine whether the buffering effect of bank concentration that was revealed in the benchmark model holds equally for banks with different balance sheet characteristics.  I try to uncover the bank characteristics, including capital level, liquidity, profitability, and size, that create this buffering effect against monetary shocks. Further, I investigate the role of corporate attributes, such as the ownership structure and the status of the public listing, in the lending channel of monetary policy transmission.

I divided the sample into various subsamples based on different bank characteristics or corporate attributes. The subsample analyses are performed using approaches similar to those in prior studies (See Hou & Wang 2013; Olivero et al. 2011a), splitting the sample for each bank-specific characteristic based on banks’ median values by country in each year. Banks with equity assets ratios above (below) the median are contained in the subsample of high- (low-)capitalisation banks. Banks with liquidity assets above (below) the median value are in the high- (low-)liquidity subsample. Banks with ROA above (below) the median level are in the high- (low-)profitability subsample. Similarly, banks with total assets above (below) the median value are categorised as large (small) banks.

The estimated results for different bank categories using the HHI-TA are reported in Table 6. The results of Tables 3 to 5 show negative and significant coefficients on the monetary policy indicator, which confirms the existence of the bank lending channel after controlling for all bank characteristics and corporate attributes, with the exception of the non-listed banks.  The results show that the reduction in loan growth is higher for the less-capitalised, less-liquid, and smaller banks. This finding is consistent with the literature showing that banks with large asset bases, which are mostly well-capitalised and highly liquid, are more resilient to monetary policy shocks (Kashyap & Stein 2000). This insulation is due mainly to their easy access to alternative sources of funds and existing capital buffers against monetary policy contractions. Hence, banks that are poorly capitalised, less liquid and with lower asset bases may have to reduce their lending capacity under contractionary monetary policy. These banks are affected the most by monetary policy shocks, as their reservable deposits are drained in monetary tightening. This is consistent with the results of Khan et al. (2016),  and Amidu & Wolfe (2013).

In addition to bank characteristics, this study also examines the roles of the different corporate attributes. Since foreign banks are excluded, the sample is divided into state-owned and private banks. In the South Asian context, state-owned banks are generally large-scale government-owned banks that operate nationwide. In these economies, it is typical for state-owned enterprises’ accounts to be maintained at state banks, and state banks generally hold a significant portion of the banking sector assets.  However, state-owned banks still maintain an arm’s-length banking relationship with the government and work toward the same profit maximisation objective as other private banks. Accordingly, the assumption of ‘profit focus’ that forms the basis of analyses of bank behaviour in the banking literature is equally applicable to state-owned banks in the South Asian context. Few studies have examined how banks’ lending behaviour varies according to a bank’s ownership structure. Bhaumik et al. (2011)conducted a study on the influence of ownership on banks’ reactions to India’s monetary policy, among public, private and foreign banks.  Their results suggested that the credit disbursal of all three types of banks are affected adversely by monetary policy tightening. However, the credit contraction was largest for foreign banks, followed by private banks. The public banks were affected the least by monetary policy shocks.  They suggested that this may have been due to the higher cost of obtaining and processing information for foreign banks, resulting in information asymmetry.

As Table 6 shows, both state-and privately-owned banks have negative and significant coefficients on the monetary policy indicator, suggesting that both types of banks are affected adversely by the monetary contraction. However, the magnitudes suggest that state-owned banks have stronger responses to monetary policy shocks than private banks. The interaction term between the monetary policy indicator and the concentration ratio is significantly positive for both state-owned and privately-owned banks. Nevertheless, the size of the interaction term suggests that an increased concentration in state-owned banks makes monetary policy transmission weaker than for privately-owned banks. The results suggest that the higher market power of the state-owned banks creates a buffering effect and reduces the effectiveness of the bank lending channel.

Table 6: Lending Channel of Monetary Policy Transmission Across Heterogeneous Bank Types

Note: This table reports panel estimates with bank and time fixed effects. The dependent variable is the loan growth rate, and the monetary policy variable is the difference in the money market rate. All control variables are lagged one period. Robust standard errors are in brackets. , *, *** denote significance at 10%, 5% and 1% respectively.

5. Conclusion

Even though there have been studies on the bank lending channel of monetary transmission, the literature on the impact of the market structure on this transmission is limited in both scope and context. This is due mainly to the lack of agreement among scholars on the measurement of market competition and the presence of conflicting evidence on the effect of banking market structure on monetary transmission. The present research examines the impact of market concentration on banks’ loan supply in South Asian banking systems. This analysis contributes to the literature by deepening the understanding of the bank lending channel in South Asia, as emerging market economies that have not been studied extensively. Hence, this study uses bank-level balance sheet and income statement data to investigate the potential impact of the market concentration on bank lending using structural measures of the market structure in South Asian economies from 2011 to 2018.

The results gathered from the analysis allow me to draw several conclusions about the bank lending channel of monetary policy in South Asian economies. The evidence confirms the existence of the bank lending channel in these economies. The structural measures of market structure show that an increased bank concentration tends to hamper the bank lending channel of monetary policy. Moreover, by examining the heterogeneous effect using different bank characteristics and corporate features, I find that banks that are large, well-capitalised, and have good liquidity positions are better insulated against monetary tightening and are affected the least. This may be due to the fact that such banks have superior access to alternative sources of funds, and hence have a capital buffer to cater to the supply of bank loans against contractionary monetary policy shocks. These results support the hypothesis that banks which are poorly-capitalised and with low liquidity cannot resist the impact of monetary contractions and therefore curtail their supply of loans.

These research questions are especially appropriate in South Asian economies, as they have undeveloped financial systems which depend heavily on banks. Accordingly, the banking sector plays a critical role in promoting developmental priorities and corporate efficiency in these emerging economies. Hence, from a policy point of view, these results indicate a pressing need for a close monitoring of developments in the banking industry that may affect the competitiveness of the banking market, such as privatization policies, the removal of entry barriers, allowing foreign banks to enter the market, etc.  Hence, this study provides insights for banking regulators which are calling for policy measures in order to tackle the adverse effects of banking sector concentration on effective monetary policy transmission.

One of the significant limitations of this study is the fact that it does not use non-structural measures as proxies for bank competition, as such information is not available. Hence, this study could be extended to use the Lerner Index or the Boone indicator for South Asian economies and then re-examine the effect of non-structural competition measures on the lending channel of monetary policy.

Availability of data and materials: The data that support the findings of this study are available on request from the corresponding author at antonette.sfernando@yahoo.com or antonette@cbsl.lk

Competing interests: The author declares no competing interests.

Funding: There is no funding for this work.

Contributions: The work is solely the contribution of the corresponding author. All errors are my own.

Declaration of Generative AI and AI-assisted Technologies: This study has not used any generative AI tools or technologies in the preparation of this manuscript.

^[1] I tried estimating Eq. (3.2) without the lagged dependent variable in a standard fixed effects panel regression framework and obtained results similar to those with the lagged dependent variable.

^[1] Under the competition-stability view, Berger et al. (2009) argued that large banks, resulting in a highly concentrated banking market, can be more diversified, take smaller risks and earn more profit. Having a smaller number of large banks in the industry, and thus regulators, enables efficient resource allocation for monitoring the stability of the banking system. In contrast, under the competition-fragility view, if these large banks believe that they are too big to fail, they are likely to undertake more risk. Such big banks also expect cover-up or assurances from policymakers. This threatens the stability, as a result of the moral hazard problem of too-big-to-fail  (Acharya et al. 2012; Boyd & De Nicolo 2005).

^[2] The identification problem will be discussed in detail in the literature review.

^[3] Bhaumik et al. (2011) argued that banks in developing countries are likely to downsize their loan portfolios more aggressively than their bond portfolios in monetary contractions, as bonds are assumed to be less risky assets (as the sovereign government is the primary issuer of bonds).

^[4] The use of bank-level data allows a better understanding of the supply-side credit channels of monetary policy transmission vs. the supply-side interest rate channel (See Olivero et al. 2011a).

^[5] Orbis presents financial statement data in a standardised format for all countries. All data are stated in millions of US dollars. In the literature, the quality of the data provided by Bureau van Dijk has been assessed as good overall.

^[6] The data outliers are eliminated following the criteria used in similar studies, for instance Olivero et al. (2011b).