The purpose of this paper is to study the determinants of Asian banks’ profitability with particular focus on the role of asset quality. This concern has been particularly important as the Basel III imposed more stringent requirements in banking regulation.
The paper uses fixed effect estimation for the panel data of the sample that consists of 947 banks from 12 Asian economies over the period of 2001-2015.
The authors find that poor asset quality (measured as impaired loans over gross loans) has a significant negative impact on banks’ profitability. Other bank-specific variables – capital adequacy, income diversification and operating inefficiency – are also important determinants. With regard to macroeconomic factors – real gross domestic product growth has most significant influence on the performance of banks.
The authors also find that the banks operating in non-advanced economies enjoy higher profit margin than banks operating in advanced economies.
Although the average asset quality in Asian banks improved over the years, governments could promote more competition, particularly in non-advanced economies. Banks in the region are recommended to diversify their income by avoiding over reliance on interest income.
Although there are prior studies that looked into asset quality, in particular with regard to the European and US experience, to the best of the authors’ knowledge there is no such study that explores cross-country Asian countries. In addition, the other primary determinants of Asian banks’ profitability are investigated. Further, the authors also looked in depth at the performance of the banks in advanced and non-advanced Asian economies.
Salike, N. and Ao, B. (2018), "Determinants of bank’s profitability: role of poor asset quality in Asia", China Finance Review International, Vol. 8 No. 2, pp. 216-231. https://doi.org/10.1108/CFRI-10-2016-0118Download as .RIS
Emerald Publishing Limited
Copyright © 2018, Emerald Publishing Limited
Bank’s primary function as a financial intermediary is to pool the funds from lenders and allocate them to borrowers, thereby making profits from the interest spread. This supposedly enhances efficient allocation of capital by channeling fund to those who have a shortage of funds from those who have a surplus. The bank loans entered into the financial statements reflect the value of the asset and measure how much financial resource is effectively transferred from banks to users.
Allen and Carletti (2010) stated that the price reflects the fundamentals of assets and can be trusted in an efficient market. However, if the price fails to do so, it may understate or overstate the fundamentals of assets and become difficult for banks to determine the value of assets. Banking is a profit-seeking industry and all the inputs and outputs need to be quantified. Hence, the price of assets held by banks can be used to measure their ability to generate future cash flow. Conventionally, one of the measures of poor asset quality is expressed as the ratio of non-performing loans or impaired loans to gross loans. While banks are able to decide the amount of loans they lend to qualified borrowers, the proportions of recoverable loans are not within the control of the management. Moreover, the recoverable loans depend on the clients’ ability to repay the principal and interest. During economic downturns, the clients’ ability to repay the principal and interest becomes weak and the pricing mechanism of an asset does not function well. In such a situation, the asset price cannot accurately reflect asset quality. Therefore, assessing the role of asset quality in the profitability of the banking sector has attracted much interest in academic research works.
Besides, regulations related to banking operations have been a point of debate among policy makers particularly after the advent of the global financial crisis (GFC) of 2007-2009. This has led to the upgrading of the Basel accord, resulting in its third installment, Basel III, to respond to the inadequacies in financial regulation revealed by the crisis. It is intended to strengthen global capital and liquidity rules with the goal of promoting a more resilient banking sector in order to improve the sector’s ability to absorb sudden shocks in the market. It also anticipates improving risk management and strengthening banks’ transparency and disclosures (BIS, 2011). In the process, it introduced several stringent regulations in both aspects of capital and liquidity requirements. For example, the minimum capital requirement was raised to 4.5 percent of risk-weighted assets and Tier 1 capital to 6 percent, the expectation for banks to have a capital conservation buffer of 2.5 percent and to maintain a leverage ratio in excess of 3 percent. Similarly, the minimum requirement for liquidity coverage ratio was increased to 100 percent and net stable funding ratio was also introduced. Following these, there are concerns how these regulations would impact lending behavior and profitability of the banks (e.g. see Ozili, 2015; Pasiouras et al., 2009).
Banks in Asia are structurally sound and appear to have stronger financial positions than the banks in Europe, the UK and Australia on all aspects of capital adequacy, liquidity and risk exposure. Nevertheless, the stringent capital and liquidity requirements have an impact on leveraging of banks that would trigger unintended reallocation of funds (Sheng, 2013). The Asian financial market is typical bank-based financial system with banking sector playing the important role of intermediary. The profitability of the banking sector was consistently positive until the GFC of 2007-2009. As can be seen in Figure 1, both the measures of profitability (return on average assets (ROAA) and return on average equity (ROAE)) were positive for Asian banks until 2007. However, these banks were quick to recover after 2009, even surpassing the pre-crisis level.
The country-specific profit distribution of the Asian banks is presented in Figure 2 for four different years. These countries have enjoyed a relatively healthy ROAA except for Japan and Malaysia over the years. Japanese banks suffered the most in the advent of the GFC. The profitability is healthy in most recent times particularly for Singapore, South Korea and Hong Kong.
Figure 3 shows the general picture of asset quality among the Asian banks, measured as average ratio of impaired loans over gross loans and average ratio of loss reserves over gross loans. It can be seen that over the years Asian banks have been effective in controlling the bad loans; in particular, the impaired loans have seen a major drop during the first half of the 2000s. The figure is stable and consistent even during the period of GFC. This shows the robust nature of the Asian banking industry.
Studies on the determinants of bank’s profitability in Asia have been undertaken to some extent in previous research works; however, there is no study that focused specially on asset quality. Further, given the heightened attention on banking regulations put forward by Basel III, it has been imperative to see if these stringent asset quality requirements have had an impact on bank’s profitability. This paper delves into such an issue by assessing the role of asset quality on profitability of Asian banks from 12 economies using the sample period of 2001-2015 that included the financial crisis years. Although there are prior studies that looked into asset quality, in particular with regard to the European and US experience, to the best of our knowledge there is no such study that explores cross-country Asian countries. In addition, the other primary determinants of Asian banks’ profitability are investigated. Further, we also looked in depth at the performance of the banks in advanced and non-advanced Asian economies. In one such study that focused on the Chinese experience, Boateng et al. (2015) examined the effect of asset quality using 111 Chinese commercial banks over the period of 2002-2012. The authors found that lower financial leverage is positively associated with overall bank performance. Among other results, they also found that foreign banks appear to have better asset quality and overall performance compared to domestic banks.
In the next section, we review the existing studies that looked into the determinants of the bank’s profitability across the world. In Section 3, we present data and the research model followed by discussions of the results in Section 4. Section 5 concludes.
2. Literature review
Previous studies on determinants of bank’s profitability have identified two set of variables: internal and external determinants from the perspective of management controllability. Internal factors are identified as capital adequacy, bank size, liquidity, market power, bank ownership and importantly, asset quality. External factors include market potential, macroeconomic condition and interest rates. While we review all the potential determinants, we pay particular attention to the role of asset quality.
2.1 Internal factors
Bank management consists of asset management, liquidity management, liability management, capital adequacy management and risk management. One of the most important determinants for bank’s profitability and the main interest of this paper is the asset quality, which is a critical variable in determining the overall condition of a bank. The quality of bank assets depends primarily on its portfolio of loans and internal credit administration mechanisms. Poor asset quality, also known as poor loan quality and normally represented by non-performing loans or impaired loans, is an important consideration in asset management and is an indicator of potential banking profitability. The theoretical consideration is that as the loans granted out by commercial banks appear in the asset side of the balance sheet, their quality would determine the credit risk for the banks. Loans are generally the most sizeable item in a bank’s assets and carry the largest potential risk to the bank’s capital account. Exposure to such risk is associated with decreased profitability (Athanasoglou et al., 2008). The excessively high level of non-performing loans could indicate poor corporate practices, lax credit administration processes and the absence of credit risk management practices. Therefore, reduction of such non- performing loans is beneficial to economic growth, eliminating the uncertainty over the banks true capital position. This would also increase the propensity to lend out more. Berger et at. (2009) and Berger et at. (2008) concluded that asset quality is associated with bank size with larger banks carrying significantly less non-performing loans, and therefore they have a better loan portfolio quality than smaller banks. In addition to bank size, the ratio of non-performing loans depends on other variables such as the borrowers’ ability to repay, the market for trading the problem assets, regulations for the resolution of non-performing loans, etc. Beck et al. (2010) used loan loss provision to proxy asset quality and concluded that there was weak evidence to show that asset quality was associated with a higher spread for private domestic banks. Kosmidou et al. (2008) found that there was no clear-cut establishment of a negative association between loans loss reserves and profitability. Nevertheless, poor asset quality is an important determinant of bank’s profitability, especially in times of recession. Bock and Demyanets (2012) found that banks’ poor asset quality is significantly linked with macroeconomic aggregates, specifically, the reduction of credit growth and worsening loan quality that could stem from lower economic growth, exchange rate depreciation, weaker terms of trade and a fall in debt-creating capital inflows. For the banks in Portugal, Garcia and Guerreiro (2016) found evidence that poor asset quality, when measured as loan loss provisions over total loans as a proxy for credit risk, has significant negative impact on profitability.
Among other internal determinants, capital adequacy, generally measured by the ratio of equity to assets, is a significant contributor to bank’s profitability. Kosmidou et al. (2008) found that well-capitalized banks (higher equity over assets ratio) had a lower risk of bankruptcy. The higher the ratio of equity over assets, the smaller the risk of insolvency that banks might face. Liquidity is the ratio of liquid assets to liquid liabilities and is a proxy for banks’ ability to repay their clients’ shot-term deposits and other funds at maturity. Kosmidou (2008) stated that a high ROAA was found to be associated with well-capitalized banks and lower cost to income ratios. In the study on the impact of liquidity on the bank’s profitability, Bordeleau and Graham (2010) concluded from the quarterly observations from 55 American bank holding companies and 10 Canadian banks from 1997 to 2009, that there is an inflection point that divides the impact of liquidity on bank’s profitability. Holding liquid assets within reasonable levels could improve bank’s profitability, but beyond the inflection point the holding of additional liquid assets diminishes banks’ financial performance. Hence, the impact of liquidity on bank’s profitability could also be ambiguous. Allen and Carletti (2010) stated that the high cost of holding liquidity may result in bank failures since the holding of liquidity increases operating costs and erodes profits whereas higher returns could have been generated by loaning out. Similar results were observed in the Indian banking sector by Sufian and Noor (2012) and Garcia and Guerreiro (2016) for Portuguese banks.
For conventional commercial banks, their income mainly consists of interest income and non-interest income. Banks make profits from the interest income that is realized through the supply of funds to private businesses. The evidence from the banking sector in Switzerland, by Dietrich and Wanzenried (2011), indicates that the share of interest income has a significant impact on profitability. Specifically, banks that are heavily dependent on interest income are not as profitable as banks whose income is more diversified. Likewise, there is a significant positive relationship between non-interest income and risk-adjusted return on assets (ROA), found by Nguyen (2012) who used data of commercial banks from 28 financially liberalized countries for the period 1997-2004.
Operating inefficiency would capture the relative inefficiencies of banks in terms of their costs and reduction of non-productive costs can result in higher profits. Dietrich and Wanzenried (2011) noted that efficient banks are more profitable than non-efficient ones. The authors examined the impact of the GFC of 2007-2009 on the profitability of commercial banks in Switzerland and concluded that the main determinants of banks’ profitability were operational efficiency, the growth of total loans, funding costs and interest income share. Moreover, loan loss provisions increased significantly during the GFC and had a negative impact on banks’ profitability.
Bank size in the form of sizeable assets is not only able to take advantage of economies of scale and earn higher incomes during prosperous periods, but is also able to soften the negative shocks during recession periods. Moreover, Haron (2004) found that the size of banks significantly influences profits based on the evidence from Islamic banks. However, Goddard et al. (2004) suggested that there was no convincing evidence to support the systematic relationship between size and operational performance although larger banks tended to further improve growth performance. This was further supported by Kosmidou (2008) who did not find a significant effect of size.
2.2 External factors
Unlike the internal factors, macroeconomic variables are the factors that are not within the control of the banks’ management. One of the frequently used variables as a measure of market potential is the growth rate of gross domestic product (GDP). Kosmidou (2008) found that GDP growth plays a substantially positive role in determining banks’ financial performance. Demand for banking services will be high during the time of upward sentiments, thereby increasing the aggregate demand. Jiang et al. (2003) found a positive relationship between real GDP growth and bank’s profitability based on a sample of banks from Hong Kong. During the economic slowdown period, uncertain global economic conditions were reflected in lower investments and export growth, which could have seriously hampered the recovery process. Therefore, higher growth rates of real GDP would result in more loans and thereby higher profitability, whereas the converse would result to lower growth.
Inflation indicates the macroeconomic condition of an economy and banks operating in countries with high inflation exhibit very high margins and cost ratios (e.g. Vittas, 1991; Flamini et al., 2009; Sastrosuwito and Suzuki, 2012). Banking activities tend to be high resulting in higher profitability during times when macroeconomic conditions are favorable. Sufian and Chong (2008) concluded that inflation has a negative impact on bank’s profitability based on empirical evidence from the Philippines from 1990 to 2005. However, Buckley (2011) argued that the net effect of inflation on bank’s profitability seems to be more complicated and ambiguous.
Interest spread, the difference between the lending interest rate and the deposit interest rate, is the nominal interest income that is generated by banks after paying depositors (Liu et al., 2010). Banks in countries with higher interest spreads tend to face lower default risks. This is because when interest spread widens, banks, as the suppliers of funds, hold assets that now have more purchasing power. Haron (2004) found that the income of Islamic banks is influenced positively by interest rates. Demirguc-Kunt and Huizinga (1999) found that higher real interest rates are associated with higher interest margins and profitability especially in developing countries. This may be because in developing countries, demand deposits frequently pay lower market interest rates.
3. Data and methodology
This study covers bank-level unbalanced panel data including 1,455 banks of 12 Asian economies from 2001 to 2015. The bank-specific variables were extracted from Orbis Bank Focus database (https://orbisbanks.bvdinfo.com/). The economies are China, Hong Kong, India, Indonesia, Japan, Malaysia, Philippines, Singapore, South Korea, Taiwan, Thailand and Vietnam. These economies were further categorized as advanced and non-advanced based on the International Monetary Fund (IMF) classification (refer to Appendix 1). The other macroeconomic variables, including growth rate of real GDP, growth rate of the CPI and interest spread, were derived from the Asian Development Bank (ADB) database.
Another consideration for the sampling of banks is the type of banks comprised in the data. The banks used in this study include: commercial banks, saving banks, cooperative banks, real estate and mortgage banks, investment banks and Islamic banks. The other categories were excluded because of less significance of these types of banks in the Asian context.
In principle, there were supposed to be 21,825 (1,455 banks ×15 years) observations but some were eliminated because of the unavailability and insufficiency of data. Several filtering processes were implemented. Specifically, those observations were excluded when banks had missing ROAA, our dependent variable. Also, observations with only one year of data as well as outliers were excluded. As a consequence, the final data set contained 10,434 observations from 947 banks (refer to Appendices 1 and 4).
For estimation purposes, the following panel relationship between the dependent variable and independent variables is generated:
Further, to look into the performance of banks operating in advanced economies vs non-advanced economics, a dummy variable, D, is added to Equation (1):
Given that the model is panel in nature, the ordinary least square may suffer from estimation bias. Therefore, we test between the use of the random and/or the fixed effects model. Under the fixed effects model, entity (bank) fixed effects and time (year) fixed effects are considered. Moreover, we suspect possible endogeneity arising from the simultaneity effect of profitability and key independent variables including asset quality. In order to deal with this, we lag the independent variables with one period.
ROAA is used as the dependent variable as a measure of bank’s profitability as it shows how efficiently the bank has utilized its assets in comparison to other banks. Our main variable of interest is asset quality (ASSQ), measured as impaired loans/gross loans. Following the discussions made in the literature review, the internal determinants used for this study are capital adequacy (CAPS) measured as equity/total assets; liquidity ratio (LIQD) measured as liquid assets/deposits and short-term borrowing; income diversification (DIV) measured as net interest revenue/average assets; and operating inefficiency (IEF) measured as cost to income ratio. Among the external determinants, we use rate of real GDP growth (GDP) to measure the market potential; inflation rate (CPI) to access the macroeconomic stability of the economy; and net interest margin (INTS) to measure the interest spread. Other variables used are ROAE and loan loss reserves/gross loans (ASSQ2) for robustness purposes.
A dummy variable (DNAD) is used to indicate non-advanced economies in the sample. Generally, in advanced and mature markets, competition among banks is intense and it is difficult for them to realize higher margins and profits (Demirguc-Kunt and Huizinga, 2000). Therefore, we hypothesize that banks in advanced economies will experience lower profit margins than banks operating in non-advanced economies. According to the IMF classification, the economies included in this study are divided into these two groups. The advanced economic entities include: Hong Kong, Japan, Singapore, South Korea and Taiwan; while the non-advanced group includes: China, India, Indonesia, Malaysia, Philippines, Thailand and Vietnam. The brief explanation of the variables, their descriptions and expected signs are provided in Table I.
4. Empirical results and discussion
4.1 Fixed effects estimation
The Hausman test was conducted to determine the choice between the fixed effects or random effects model. The fixed effects model explores the relationship between predictor or independent variables and outcome variables within an entity with each entity having its own individual characteristics. Unlike the fixed effects model, the random effects model assumes that variation across entities is random and uncorrelated with predictors included in the model. The null hypothesis of the Hausman test is that the difference in coefficients is not systematic (i.e. the preferred model is random) against the alternative hypothesis that fixed effects model is preferable. The test result is shown in Table II (Note 1). The probability is 0.000 and is far smaller than 0.05, therefore, the null hypothesis is rejected and the fixed effects model is preferred.
Further, the estimated coefficients are obtained by using clustered standard errors because of the statistical significance of the result of the Breusch-Pagan/Cook-Weisberg test (null hypothesis of constant variance) as shown in Table II (Note 2). The lower probability 0.00 indicates the presence of heteroskedasticity and the null hypothesis should be rejected.
Table II presents the results of regression for the different specifications.
Firstly, our main variable of interest, poor asset quality (ASSQ) is taken as the single determinant of bank’s profitability and reported in column 1. It has significant negative impact on banks’ returns although the magnitude is small. As it is defined as the ratio of impaired loans to gross loans, this result indicates that a 1 percent increase in impaired loans to gross loans would reduce the ROAA by 0.007 percent. When bank managers decide to accept a risky policy and lower the requirements needed to obtain a loan, for example, by providing loans to clients who are unable to afford sufficient charges or provide liquid assets as collateral, then the probability of default will increase once the customers’ financial situations worsen. However, if the management implements a conservative credit policy and attempt to minimize impaired loans through the setting of strict filtering rules and proper investigation of the customers’ background and capacity to repay, the probability of non-performing loans could be reduced.
In column 2, other bank-specific variables are considered. The results indicate that the impact of poor asset quality (ASSQ) on bank’s profitability is still negative and highly significant. The coefficient of capital adequacy (CAPS) is positive and significant, meaning banks with strong capital arrangements experience higher returns. Liquidity (LIQD) and income diversification (DIV) both have a positive association with banks’ profitability. Interest income is the bank’s core revenue generation activity and its higher value would indicate that banks are less diversified. This result is in contrast to our expectation. A possible explanation for this is that in Asia, as banks operate more in the traditional model, interest income dominates the total income structure. The banks’ operating inefficiency (IEF) has an adverse effect on profitability. For a profit-oriented business, any inefficiency in operation (larger proportion of expenses to total revenue) will erode the company’s financial performance.
In column 3, we add macroeconomic variables: growth rates of real GDP, growth rates of consumer price index (CPI) and interest spread (INTS). Poor asset quality still remains negative and is a highly significant determinant of bank’s profitability.
The growth rate of GDP positively contributes to banks’ financial performance, consistent with the literature. Both inflation and interest spread are found to be not significant.
In column 4, we report the result by including the dummy variable for non-advanced economies (DNAD); DNAD equals to 1 if the bank is operating in a non-advanced economy, 0 otherwise. The estimation technique used for this specification is the random effects estimation. It is clear that the coefficient of the dummy variable is positive and significant. It reveals the notable difference indicating that banks operating in non-advanced economies do better than those operating in advanced economies in terms of profitability. This could be explained due to competition factor. In the advanced economy, the banking industry has been growing for a long time and customers can freely transfer from their current bank to another so as to enjoy better banking services. This is because in a mature market, the customers’ transferring cost is lower and they do not need to stay loyal to one bank and sacrifice the convenience and flexibility that can be provided by another bank. Additionally, the products and service provided by one bank can be substituted by others in advanced economies; this threat forces the banks to reduce internal non-productive overheads and external charges.
In all four specifications, our main variable of interest, poor asset quality (ASSQ), maintained its negative sign and significant at 1 percent. The magnitude is rather small ranging from 0.007 to 0.013 percent, nevertheless the results indicate the adverse effect of poor asset quality in the bank’s profitability.
4.2 Entity and year fixed effects estimation
Assuming that economies might face time invariant effects, we next conduct the regression with entity and time fixed effects. The results are reported in Table III followed by the discussion.
We present the results of the random effects model in column 1 for the purpose of comparison. The random effects model does not require the specification of individual characteristics that may influence the predictor variables. In column 2, we replicate the results of entity fixed effects model (as reported in Table II, column 3), again for the purpose of comparison.
In the last column, we present the results of entity (bank) and time (year) fixed effects. The assumption made is that the performance of banks is affected not only by entity-specific factors but also by time invariant factors. Our main variable of interest, poor asset quality (ASSQ), is still negative and highly significant at 1 percent. There has been a slight increase in the magnitude to 0.10 after taking time effects into account. It indicates that if the ratio of impaired loans to gross loans increase by 1 percent, the ROAA would go down by almost 0.01 percent. The joint test for time dummy variables is also significant at 1 percent.
With regard to other variables, there are no major changes in both the bank-specific and macroeconomic variables. All these variables keep the same sign as entity (bank) fixed effects results and there is also not much difference in magnitude. Interest spread has a change in sign but is not significant.
Overall, with the above results in hand, we find the evidence that poor asset quality (ASSQ) plays an important role that has a negative effect on profitability of Asian banks. The results pertaining to the asset quality variables are consistent over different models, considering for entity and/or time fixed effects. It shows that with the increase in the loans that banks fail to collect from the borrowers, termed as impaired loans, there are direct negative consequences on a bank’s ROA. The absolute impact in this case of Asian banks is somewhat minimal. In general, 100 percent increment in impaired loans would result in just 1 percent reduction in bank’s profitability. This further shows that Asian banks are structurally sound. Other bank-specific determinant variables also pose significant effects whereas GDP is the main macroeconomic variable that produces significant results.
4.3 Robustness test
We conducted several robustness tests, in particular, to deal with two aspects of analysis. First, to control the bias on sample due to the large sample size of one particular country, Japan. Second, to examine the use of the dependent variable and the main variable of interest. The results are reported in Table IV.
In column 1, we reproduce the fixed effects estimation results of Table II (column 3) to serve as a benchmark comparison with the other results. In our sample, the number of observations from the banks of Japan outnumbers others and account for about 70 percent of the total banks (refer to Appendix 4). Hence, we doubt if the results in column 1 are characterized by one single country. To eliminate this, all the banks operating in Japan are excluded from the model while all the independent variables and the dependent variable remain the same, as shown in the column 2. The output indicates that all the bank-specific variables and macroeconomic variables included in this paper are still important determinants of banks’ profitability and such a result is consistent with the findings in column 1. They imply that the previous conclusions about the relationships between outcome variables and bank-specific and macroeconomic condition-related predictor variables are robust in nature from the perspective of country composition, particularly with regard to significant negative influence of poor asset quality.
Next, we examine the relationship between banks’ profitability and poor asset quality by changing the dependent variable to ROAE. Comparing models in columns 1 and 3, all determinants are consistent and are still statistically significant except for inflation, which is now negative and significant. In particular, poor asset quality, our main variable of interest, is still an important factor to determine Asian banks’ financial performance.
We also examine the relationship between banks’ profitability and poor asset quality by changing our variable of interest, poor asset quality. The results reported in column 4 use the banks’ asset quality as loan loss reserves per gross loans (ASSQ2). We find that the new measure of asset quality variable is also negative and highly significant thereby confirming our main results. Lastly, in column 5, we report results by adding lagged dependent variable as one of the independent variables in the main regression. The assumption that this year’s profitability is affected by last year’s performance is confirmed by positive significant results. Moreover, signs and significance of other variables are consistent with the previous results and expectations.
Overall, we find that poor asset quality is an important determinant that contributes to the banks’ financial performance despite tailoring the sample size through the exclusion of banks from Japan, the changing of the dependent variable and the use of different measures of variables. The most encouraging result is that in all the robustness results, our variable of interest, poor asset quality, appears consistently with a negative sign and is highly significant. These findings validate the robustness of our original results.
5. Conclusions and recommendations
Profitability of banks is determined by internal and external factors from the perspective of control of bank management. Among one of the most important internal factors is the quality of assets. It is expected that the poorer the quality of the asset, the more adverse effect it will have on profitability.
Furthermore, after the GFC of 2007-2009 and the introduction of stringent requirements by Basel III, there have been consequences on the lending behavior and profitability of the banks operating in the region. Policy makers are concerned that the regulations have resulted in the reallocation of loans to different sectors.
In this paper, we looked into the role of asset quality in the profitability of Asian banks. We found that poor asset quality is significantly and negatively associated with banks’ financial performance. This implies that any increase in poor asset quality will indicate a lower return for the bank because more loans are likely to be provisioned or directly written-off if this ratio gets bigger. For banks following the prudent principle of the International Financial Reporting Standards, the assets held by the borrowers in the form of bank loans cannot be overstated in the statements of financial position. Therefore, the probability of loan recoveries does matter. When indicators suggest that some loans will not be recovered, banks must record such a loss and charge it to the income statement. Eventually, such banks suffer from underperformance, which is reflected as lower returns on the assets or equity employed during that period.
For regulators in non-advanced economies, further efforts are needed to build a well-functioning capital market where the banking industry plays a crucial role. One of the ways to develop the banking industry in less advanced economies is to make the domestic banks open to more competition and to introduce more financial institutions. Temporary protection regulations may help domestic banks earn higher profits in the short term; however, these protections are not likely to guarantee success in the long run. Another benefit of having foreign financial institutions is to bring in more managerial and product skills. Second, there are other bank-specific variables that are highly likely to contribute to a better financial performance, in particular, capital adequacy, income diversification and operating inefficiency. This finding is consistent with earlier research. In addition, all these factors are related to the banks’ operation, specifically, these determinants are largely within the control of the management. Hence, any decisions about capital adequacy, liquidity, channels of income generation and operating efficiency should be made deliberately by the management to strike a balance between risk and return in the face of asymmetric information. Third, real GDP is found to be an important macroeconomic factor in the performance of bank’s profitability; higher GDP growth is associated with higher profits. Compared to the bank-specific factors, these determinants are not within the control of bank management. Therefore, it is important for banks to choose feasible strategies under such circumstances.
Finally, there exists a notable difference in profitability for banks operating in advanced vs non-advanced economies. This finding is consistent with the prediction that banks operating in advanced economies are likely to experience lower profit margins as intensified competition weakens the interest spread. The banks in Asia may consider diversifying their income by avoiding over reliance on interest income, in the wake of growing competition. Furthermore, the threat of new entrant banks forces many existing banks to acquire or merge with others to achieve economies of scale. To survive and remain competitive in the industry, most banks should choose either cost leadership or to focus on niche markets.
Variable definition and notation
|Dependent variable||Bank’s profitability||Return on average assets||ROAA|
|Internal determinants||Poor asset quality||Impaired loans/Gross loans||ASSQ||−|
|Capital adequacy||Equity/Total assets||CAPS||+/−|
|Liquidity ratio||Liquid assets/Deposits and short-term borrowing||LIQD||+/−|
|Income diversification||Net interest revenue/Average assets||DIV||−|
|External determinants||Market potential||Growth rate of real GDP||GDP||+|
|Macroeconomic condition||Growth rate of CPI||CPI||+/−|
|Interest spread||Net interest margin||INTS||+|
|Country group||Dummy variable: 1=non-advanced countries; 0=advanced countries||DNAD||+|
Notes: Refer Appendix 2 for descriptive statistics of the variables. All variables are in percentage except for the dummy variable
Panel regression with fixed effects
|Dependent variable: return on average assets||(1)||(2)||(3)||(4)|
|Poor asset quality||−0.007*** (−2.92)||−0.008*** (−3.18)||−0.009*** (−3.37)||−0.013*** (−6.31)|
|Capital adequacy||0.028*** (3.88)||0.027*** (3.83)||0.025*** (5.31)|
|Liquidity ratio||0.003** (2.35)||0.003** (2.37)||0.002** (2.25)|
|Income diversification||0.021*** (2.81)||0.024*** (3.03)||0.024*** (3.20)|
|Inefficiency||−0.004*** (−5.83)||−0.004*** (−5.10)||−0.006*** (−8.35)|
|Growth rate of real GDP||0.012*** (6.08)||0.011*** (5.4)|
|Growth rate of CPI||−0.004 (−1.04)||−0.004 (−1.19)|
|Interest spread||−0.016 (−1.08)||−0.014 (−0.97)|
|Dummy for non-advanced country||0.378*** (3.83)|
|Constant||0.409*** (25.61)||0.412*** (4.91)||0.397*** (4.51)||0.575*** (6.96)|
|No. of observations||8,934||8,563||8,563||8,563|
Notes: 1. Hausman test results; H0: difference in coefficients is not systematic; χ2=260.5; Prob>χ2=0.00002. Breusch-Pagan/Cook-Weisberg test for heteroskedasticity; H0: constant variance; χ2=723.91; Prob>χ2=0.00003. Standard errors clustered at bank level4. Economy dummy variables are also used for random effects model in column 4Numbers in parentheses indicate t-statistics. **,***Statistically significant at 5 and 1 percent levels, respectively
Panel regression with entity and year fixed effects
|Dependent variable: return on average assets||(1)||(2)||(3)|
|Poor asset quality||−0.013*** (−8.53)||−0.009*** (−3.37)||−0.010*** (−3.09)|
|Capital adequacy||0.025*** (9.17)||0.027*** (3.83)||0.028*** (3.76)|
|Liquidity ratio||0.002*** (3.04)||0.003** (2.37)||0.002* (1.90)|
|Income diversification||0.024*** (6.15)||0.024*** (3.03)||0.043*** (5.09)|
|Inefficiency||−0.006*** (−15.11)||−0.004*** (−5.10)||−0.004*** (−5.14)|
|Growth rate of real GDP||0.011*** (4.48)||0.012*** (6.08)||0.012** (2.57)|
|Growth rate of CPI||−0.004 (−1.41)||−0.004 (−1.04)||−0.008 (−1.63)|
|Interest spread||−0.014 (−1.27)||−0.016 (−1.08)||0.013 (0.67)|
|Constant||0.952*** (10.61)||0.397*** (4.51)||0.318*** (3.55)|
|No. of observations||8,563||8,563||8,563|
|Bank fixed effects||No||Yes||Yes|
|Time fixed effects||No||No||Yes|
Notes: 1. Joint test for time dummies: F (13, 911)=24.62; Prob>F=0.00002. Standard errors clustered at bank level3. Economy dummy variables are also used for random effects model in column 1 Numbers in parentheses indicate t-statistics. *,**,***Statistically significant at 10, 5 and 1 percent levels, respectively
Panel regressions for robustness check
|Dependent variable||Return on average assets||Return on average assets||Return on average equity||Return on average assets||Return on average assets|
|Poor asset quality||−0.009*** (−3.37)||−0.012** (−2.17)||−0.145** (−2.56)||−0.006** (−2.20)|
|Capital adequacy||0.027*** (3.83)||0.041*** (5.1)||0.230** (2.42)||0.021** (2.51)||0.020*** (2.93)|
|Liquidity ratio||0.003** (2.37)||0.001 (0.83)||0.057*** (2.62)||0.002* (1.89)||0.003** (2.43)|
|Income diversification||0.024*** (3.03)||0.024*** (2.88)||0.543*** (5.30)||0.020** (2.45)||0.021*** (2.94)|
|Inefficiency||−0.004*** (−5.10)||−0.011*** (−4.79)||−0.078*** (−2.83)||−0.004*** (−5.06)||−0.002** (−2.50)|
|Growth rate of real GDP||0.012*** (6.08)||0.017*** (3.78)||0.241*** (5.63)||0.010*** (4.72)||0.011*** (5.75)|
|Growth rate of CPI||−0.004 (−1.04)||−0.003 (−0.60)||−0.151** (−2.26)||0.002 (0.56)||−0.005 (−1.37)|
|Interest spread||−0.016 (−1.08)||0.012 (−0.57)||−0.367 (−1.54)||−0.032** (−2.09)||−0.011 (−0.75)|
|Loss reserves per gross loans||−0.018** (−2.48)|
|Return on average assets (lagged)||0.125*** (5.02)|
|Constant||0.397*** (4.51)||0.938*** (5.72)||10.026*** (4.47)||0.394*** (3.97)||0.246*** (2.88)|
|No. of observations||8,563||2,267||8,560||8,563||8,563|
Notes: 1. In column 1, banks from all 12 economies are included2. In column 2, banks from Japan are excluded from the regression3. In column 3, the dependent variable is taken as return on average equity (ROAE)4. In column 4, the banks’ poor asset quality is measured as loan loss reserves/gross loans5. In column 5, lagged dependent variable is added as one of the independent variables6. Standard errors clustered at bank levelNumbers in parentheses indicate t-statistics. *,**,***Statistically significant at 10, 5 and 1 percent levels, respectively
Classification of economies and number of banks
|No. of banks|
|Country group||Before filtering||After filtering|
Notes: Included banks: commercial banks, saving banks, cooperative banks, real estate and mortgage banks, investment banks and Islamic banks (Source: Orbis Bank Focus database); for the classification of advanced economies and non-advanced economies (Source: International Monetary Fund (IMF))
|Return on average assets||10,428||0.37||0.68||−6.51||13.79|
|Poor asset quality||10,313||6.33||4.94||0.00||37.35|
|Growth rate of real GDP||10,434||2.42||3.54||−5.53||15.24|
|Growth rate of CPI||10,434||1.33||2.91||−1.37||23.09|
|Return on average equity||10,421||4.76||11.50||−298.18||67.25|
Correlation coefficient of variables
Frequency of banks
Grubbs test was used to identify the outliers.
Since DNAD is a dummy variable, the fixed effects model could not be implemented.
Allen, F. and Carletti, E. (2010), “An overview of the crisis: causes, consequences, and solutions”, International Review of Finance, Vol. 10 No. 1, pp. 1-26.
Athanasoglou, P.P., Brissimis, S.N. and Delis, M.D. (2008), “Bank-specific, industry specific and macroeconomic determinants of bank profitability”, Journal of International Financial Markets, Institutions & Money, Vol. 18 No. 2, pp. 121-136.
Beck, T., Cull, R., Fuchs, M., Getenga, J., Gatere, P., Randa, J. and Trandafir, M. (2010), “Banking sector stability, efficiency, and outreach in Kenya”, Policy Research Working Paper No. 5442, The World Bank, Washington, DC.
Berger, A.N., Klapper, L.F. and Turk-Ariss, R. (2009), “Bank competition and financial stability”, Journal of Financial Services Research, Vol. 35 No. 2, pp. 99-118.
Berger, A.N., Young, R.D., Flannery, M.J., Lee, D. and Öztekin, Ö. (2008), “How do large banking organizations manage their capital ratios”, Journal of Financial Services Research, Vol. 34 Nos 2/3, pp. 123-149.
BIS (2011), Basel III: A Global Regulatory Framework for More Resilient Banks and Banking Systems, Basel Committee on Banking Supervision and Bank for International Settlements, Basel.
Boateng, A., Huang, W. and Kufuor, N.K. (2015), “Commercial bank ownership and performance in China”, Applied Economics, Vol. 47 No. 49, pp. 5320-5336.
Bock, R.D. and Demyanets, A. (2012), “Bank asset quality in emerging markets: determinants and spillovers”, Working Paper No. Wp/12/71, International Monetary Fund, Washington, DC.
Bordeleau, E. and Graham, C. (2010), “The impact of liquidity on bank profitability”, Bank of Canada Working Paper No. 2010-38, Ottawa.
Buckley, A. (2011), Financial Crisis: Causes, Context and Consequences, Financial Times Prentice Hall, Edinburgh.
Demirguc-Kunt, A. and Huizinga, H. (1999), “Determinants of commercial bank interest margins and profitability: some international evidence”, The World Bank Economic Review, Vol. 13 No. 2, pp. 379-408.
Demirguc-Kunt, A. and Huizinga, H. (2000), “Financial structure and bank profitability”, Policy Research Working Paper No. 2430, The World Bank, Washington, DC.
Dietrich, A. and Wanzenried, G. (2011), “Determinants of bank profitability before and during the crisis: evidence from Switzerland”, Journal of International Financial Markets, Institutions & Money, Vol. 21 No. 3, pp. 307-327.
Flamini, V., McDonald, C. and Schumacher, L. (2009), “The determinants of commercial bank profitability in Sub-Saharan Africa”, Working Paper No. WP/09/15, International Monetary Fund, Washington, DC.
Garcia, M.T.M. and Guerreiro, J.P.S.M. (2016), “Internal and external determinants of banks’ profitability: the Portuguese case”, Journal of Economic Studies, Vol. 43 No. 1, pp. 90-107.
Goddard, J., Molyneux, P. and Wilson, J.O.S. (2004), “The profitability of European banks: a cross- sectional and dynamic panel analysis”, The Manchester School, Vol. 72 No. 3, pp. 363-381.
Haron, S. (2004), “Determinants of Islamic bank profitability”, The Global Journal of Finance and Economics, Vol. 1 No. 1, pp. 11-33.
Jiang, G., Tang, N., Law, E. and Sze, A. (2003), “The profitability of the banking sector in Hong Kong”, Hong Kong Monetary Authority Quarterly Bulletin, Vol. 3 No. 36, pp. 5-14.
Kosmidou, K. (2008), “The determinants of banks’ profits in Greece during the period of EU financial integration”, Managerial Finance, Vol. 34 No. 3, pp. 146-159.
Kosmidou, K., Tanna, S. and Pasiouras, F. (2008), “Determinants of profitability of domestic UK commercial banks: panel evidence from the period 1995-2002”, Economics, Finance and Accounting Applied Research Working Paper Series No. RP08-4, Coventry University, Coventry.
Liu, H., Molyneux, P. and Nguyen, L.H. (2010), “Competition and risk in the South East Asian commercial banking”, Applied Economics, Vol. 44 No. 28, pp. 3627-3644.
Nguyen, J. (2012), “The relationship between net interest margin and noninterest income using a system estimation approach”, Journal of Banking & Finance, Vol. 36 No. 9, pp. 2429-2437.
Ozili, P.K. (2015), “Determinants of Bank profitability and Basel capital regulation: empirical evidence from Nigeria”, Research Journal of Finance and Accounting, Vol. 6 No. 2, pp. 124-131.
Pasiouras, F., Tanna, S. and Zopounidis, C. (2009), “The impact of banking regulations on banks’ cost and profit efficiency: cross-country evidence”, International Review of Financial Analysis, Vol. 18 No. 5, pp. 294-302.
Sastrosuwito, S. and Suzuki, Y. (2012), “The determinants of post-crisis Indonesian banking system profitability”, Economics and Finance Review, Vol. 1 No. 11, pp. 48-57.
Sheng, A. (2013), “Basel III and Asia”, Finance working paper, Fung Global Institute, Hong Kong.
Sufian, F. and Chong, R.R. (2008), “Determinants of Bank profitability in a developing economy: empirical evidence from the Philippines”, Asian Academy of Management of Journal of Accounting and Finance, Vol. 4 No. 2, pp. 91-112.
Sufian, F. and Noor, M.A.N.M. (2012), “Determinants of Bank performance in a developing economy: does bank origins matters”, Global Business Review, Vol. 13 No. 1, pp. 1-23.
Vittas, D. (1991), “Measuring commercial Bank efficiency: use and misuse of bank operating ratios”, Working Paper Series No. 806, The World Bank, Washington, DC.
The authors would like to thank Yilin Jolene Tan for her able research assistance and Shafeena Taylor-Cross for proof reading. The authors are also grateful to two anonymous referees and the editors for their valuable and insightful comments, which improved the quality of the paper.