International Journal of Islamic Economics and Finance (IJIEF) Vol. 6(1), January 2023, pages 1-20 The Cyclical Behavior of Capital Buffer of Indonesian Islamic and Conventional Banks Nur Atikah1, Akhmad Akbar Susamto2, Samsubar Saleh3, Ficrat Abdic4 Corresponding email: tikanasaa@gmail.com Article History Received: November 18th, 2021 Revised: June 13th, 2022 June 16th, 2022 July 2nd, 2022 December 1st, 2022 December 12nd, 2022 December 29th, 2022 Accepted: January 3rd, 2023 Abstract This paper is aimed at analyzing the cyclical behavior of capital buffer of Islamic and conventional banks in Indonesia. More specifically, this paper has three objectives. First, to test whether capital buffer in Indonesian banking industry as a whole is countercyclical or procyclical. Second, to test whether there is a difference in the level of capital buffer of Islamic banks as compared to the level of capital buffer of conventional banks. Third, to test whether there is a difference in the cyclicality of capital buffer of Islamic and conventional banks. The analysis in this paper is conducted using the standard dynamic system generalized method of moments (system GMM) regressions and includes a panel of 108 banks over the period between 2004 and 2019. From the results, it can be concluded that the capital buffer of Islamic and conventional banks in Indonesia is procyclical. From the results, it can also be concluded that no difference exists in the level of capital buffer of Islamic banks as compared to conventional banks and in the cyclicality of capital buffer in Islamic and conventional banking. If the countercyclical capital buffer is achieved, a policy measure to alter the cyclical behavior of capital buffer of Islamic and conventional banks in Indonesia therefore is a must. Such policy measure needs not to be specified for Islamic or conventional banking industry. Keywords: Basel III; Capital buffer; Cyclicality; Bank; Indonesia JEL Classification : G21; G28 Type of paper: Research Paper @ IJIEF 2023 published by Universitas Muhammadiyah Yogyakarta, Indonesia DOI: https://doi.org/10.18196/ijief.v6i1.13154 Web: https://journal.umy.ac.id/index.php/ijief/article/view/13154 Citation: Atikah, N., Susamto, A. A., Saleh, S., & Abdic, F. (2023). The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks. International Journal of Islamic Economics and Finance (IJIEF), 6(1), 1-20. DOI: https://doi.org/10.18196/ijief.v6i1.13154. 1,2,3,4 Universitas Gadjah Mada, Indonesia mailto:tikanasaa@gmail.com https://doi.org/10.18196/ijief.v6i1.13 https://doi.org/10.18196/ijief.v6i1. https://crossmark.crossref.org/dialog/?doi=10.18196/ijief.v6i1.13154&domain=pdf Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 2 I. Introduction Drawing attentions on the Global financial crisis that took place in 2007–2009, the Basel Committee on Banking Supervision (BCBS) has introduced a regulatory framework which emphasizes the need for countercyclical capital buffer in addition to the minimum capital requirements. The framework, called Basel III, seeks to increase the ability of banks to endure unexpected shocks and to maintain banking stability. Capital buffer concerns the holding of capital above the regulatory threshold. In the Basel III framework, it consists of two components, namely capital conservation buffer (CCoB) which amounts 2.5 percent of total risk-weighted assets and countercyclical capital buffer (CCyB) which can vary from countries to countries. The capital conservation buffer is designed to guarantee that banks will raise their capital buffer over non-recession periods. Meanwhile, the countercyclical capital buffer is designed to guarantee that the ups-and- downs movement of capital buffer is proper along the business cycle (BCSB, 2010). Countercyclical capital buffer exists when the relationship between capital buffer and the business cycle is positive. That is, banks increase capital buffer during economic expansions and consume the buffer to cover losses during economic downturns. By contrast, capital buffer is said to be procyclical when its relationship to the business cycle is negative. In this regard, banks lower capital buffer at the time of economic upturns and increase capital buffer at the time of economic recessions. The question of whether capital buffer in conventional banking industry is countercyclical or procyclical has been subject to many studies. These include studies published before the release of the Basel III framework, such as Ayuso et al. (2004), Lindquist (2004), Bikker & Metzemakers (2005), Marcucci & Quagliariello (2008), and Jokipii & Milne (2008). Studies that were published after the release of the Basel III framework include Stolz & Wedow (2011), Prasetyantoko & Soedarmono (2010), Deriantino (2011), Coffinet et al. (2012), Guidara et al. (2013) Shim (2013), Carvallo et al. (2015) and Huang & Xiong (2015). The number of studies examining the cyclical behavior of capital buffer in Islamic banking industry is relatively limited (Maatoug et al., 2019; Maharani & Setiyono, 2018). Maharani & Setiyono (2018) conduct an analysis investigating the determinants of capital buffer of Islamic and conventional banks in the Southeast Asian (ASEAN) and Middle-East and North African (MENA) countries. Focusing on the case of MENA countries, Maatoug et al. (2019) investigate whether the capital buffer of Islamic and conventional banks is countercyclical or procyclical. Other studies (Bougatef & Korbi, 2018; Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 3 Daher et al., 2015), while examining capital buffer in Islamic as compared to conventional banking industry, focus more on the role of risk exposure rather than the effect of the business cycle. This paper aims to analyze the cyclical behavior of capital buffer of Islamic and conventional banks in Indonesia. More specifically, the objectives of this paper are as follows. First, to test whether capital buffer in Indonesian banking industry is countercyclical or procyclical. Second, to test whether there is a difference in the level of capital buffer of Islamic banks as compared to the level of capital buffer of conventional banks. Third, to test whether there is a difference in the cyclicality of capital buffer of Islamic and conventional banks. This paper contributes to literature in several ways. First, this paper expands on the previous work by Prasetyantoko & Soedarmono (2010) that examines capital buffer in Indonesia –one of the largest economies in the world whose capital buffer cyclical behavior remains understudied. This paper is, however, different in that it includes Islamic and conventional banks in the sample. Second, this paper adds to the previous works by Maharani & Setiyono (2018) and Maatoug et al. (2019). However, rather than limiting itself to the analysis of capital buffer determinants or to the relationship between capital buffer and the business cycle, this paper compares the level of capital buffer of Islamic and conventional banks and looks at the possible difference between Islamic and conventional banks in term of capital buffer cyclicality. The remainder of this paper is organized as follows. Section II discusses the theoretical framework of capital buffer cyclical behavior and earlier empirical findings. Section III describes the data, regression models and estimation methods. Section IV summarizes and discusses the results of the analysis. Section V concludes and proposes recommendations. II. Literature Review 2.1 Background Theory Holding a higher level of capital can be costly for a bank. The most notable costs stem from the fact that it reduces the bank’s assets growth and lowers the bank’s profitability, particularly in a market that is relatively competitive. Other costs stem from the fact that many investors prefer investments that are liquid and safe with moderate returns such as providing debt to a bank rather than investments that, despite offering potential high returns, are illiquid and risky such as buying a bank’s equity. From the bank’s point of view, it is therefore easier and cheaper to obtain debt than equity financing. Moreover, in many countries, debt has been given a more favorable tax Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 4 treatment than equity (De Mooij, 2011). Debt in the form of third-party deposits has also been benefitted from deposit insurance schemes, connoting the disadvantage of holding a higher level of capital (Kane, 1989). However, there could be different reasons why a bank may hold a higher level of capital and maintain its capital ratio above the regulatory threshold. For example, as presented by Huang & Xiong (2015), there are three reasons for a bank to maintain its capital buffer. First, to minimize potential asymmetric information between the bank and its depositors. Having a higher level of capital will signal the bank's financial health and build up the depositors' confidence in the bank's ability to absorb unexpected shocks (Jackson et al., 1999). Second, a bank may hold a higher level of capital to be precautionary against the risk of failure. Having a higher level of capital will allow a bank to put on guard against unexpected shocks and reduce the probability of bankruptcy, particularly by increasing its ability to cover losses during economic downturns (Nier & Baumann, 2006). Having a higher level of capital will also allow a bank to avoid the costs of failure, which include among others, the loss of charter value, the loss of the bank's reputation and goodwill, and the loss of resources during the bankruptcy process (see, for instance, Acharya (1996)) Third, to follow Ayuso et al. (2004), Milne (2004), and Milne & Whalley (2011), a bank may hold a higher level of capital to reduce the probability of violating the minimum capital requirements. By having a capital buffer, a bank protects itself against pricey adjustment costs that may arise due to regulatory threshold violations. For example, in the presence of information asymmetries, capital level adjustment via sudden equity issues may convey a negative signal to the market that the bank's stock prices are above the actual share values. Thus, in addition to the adjustment costs, pure transaction costs may include significant drops in the importance of the bank's stocks (see, for instance, Myers & Majluf (1984), Winter (1994), and McNally (1999)). Meanwhile, capital level adjustment via asset reductions may force a bank to sell profitable assets at prices far below their values or renounce potential projects that would otherwise have been financed. Formally, the costs of capital level adjustment and the effect that the business cycle has on capital buffer can be modeled within a partial adjustment framework. To follow Ayuso et al. (2004), Jokipii & Milne (2008), Carvalho & Castro (2015) and Maatoug et al. (2019), the framework is represented by ∆Buf𝑖,𝑡 = β(Buf𝑖,𝑡 ∗ − Buf𝑖,𝑡−1) (1) where ∆ denotes the difference or the change operator, Buf𝑖,𝑡 and Buf𝑖,𝑡−1 denote the actual capital buffer of bank 𝑖 at time 𝑡 and time 𝑡 − 1 respectively, Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 5 Buf𝑖,𝑡 ∗ denotes the optimum capital buffer of the same bank and β denotes the speed of adjustment of the bank’s actual capital buffer towards its optimum level. By adding Buf𝑖,𝑡−1 to both sides, the previous equation can be rewritten as Buf𝑖,𝑡 = βBuf𝑖,𝑡 ∗ + (1 − β)Buf𝑖,𝑡−1 (2) The optimum capital buffer level (BUF∗i,t) is not observable. However, it may depend on the business cycle. Here, the relationship between capital buffer and the business cycle can be explained using at least two different approaches. First, as suggested by Stolz & Wedow (2011), the choice between loan portfolio and capital buffer changes along with the business cycle. In upturn economic conditions, loan risk is relatively low. This induces a bank to increase its loan disbursements at the expense of capital buffer. By the same token, in downturn economic conditions, as loan risk increases, a bank will reduce its loan disbursements and increase capital buffer. Second, as stated by Amato & Furfine (2004), the choice between loan portfolio and capital buffer changes along with loan quality. In upturn economic conditions, loan quality is relatively good and the possibility of default on payments is low. This induces a bank to increase its loan disbursements at the expense of capital buffer. By the same token, in downturn economic conditions, as loan quality decreases, a bank will reduce its loan disbursements and increase capital buffer. Similar to its conventional counterpart, an Islamic bank may hold a high level of capital to reduce the probability of violating the minimum capital requirements. While an Islamic bank is not allowed to pay or to charge interest and instead performs intermediation function using Islamic financial contracts (Askari et al., 2015; Habib, 2018), its optimum capital buffer may remain dependent on the business cycle. In brief, the Islamic financial contracts used by an Islamic bank consists of mudarabah, musharakah, murabahah and ijarah. Mudarabah is a business partnership contract, which one partner as an investor provides capital to its partner through a profit-sharing arrangement agreed before. Profits will be shared between the investor and its partner according to the agreed arrangement, while losses will be borne by the investor. Mudarabah can be applied for both funding and financing operations. Mudarabah that is applied for an Islamic bank’s funding operation does not guarantee the funds deposited by depositors. Profits obtained will be shared between the bank and the depositors, while losses will be assumed by the depositors. From the bank’s perspective, this theoretically implies no risk. Meanwhile, mudarabah that is applied for financing operation is risky and may lead to an increase in the probability of a bank violating the minimum capital requirements. The risk Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 6 to the bank arises from the potential of failure payments from its partners due to business losses. The risk is low in upturn economic conditions, inducing a bank to increase its financing disbursements and reduce capital buffer. Congruently, the risk is high in downturn economic conditions, inducing the bank to reduce its financing disbursements and increase capital buffer. Musharakah, murabahah and ijarah are applied mostly for financing operations. Musharakah is a business partnership contract in which two or more parties provide both capital and managerial ability in a pre-agreed agreement. All partners will share the profits according to the agreement, while losses will be borne to all partners according to the level of the capital provided. Murabahah is a selling contract in which a bank buys a spesific investment goods and sells it to the costumers with a payments arrangement and agreed margin at a future date. Lastly, ijarah is a rent contract in which a bank provide a services of product on leases to its customers and the customers pay for rent over a specified period. Similar to mudarabah that is applied for financing operation, musharakah contract bears risk that arises from the possobility of the failure payments from its partners due to business losses. Murabahah and ijarah bring about risk, particularly from the possibility of the failure payments from bank’s customers to pay the installments rents according to the agreed arrangements. The risk is, once again, low in upturn economic conditions and high in downturn economic conditions. This induces the bank to adjust its financing disbursements to the detriment of capital buffer. Theoretically, the risk to an Islamic bank in mudarabah and musharakah contracts can be higher than the risk to a conventional bank that performs intermediation function using an interest-based loan contract. This is particularly true in the presence of both asymmetric information and adverse selection (Khan, 2010). The risk in murabahah dan ijarah, while not exactly the same, is quite close to the risk in an interest-based loan contract. A priori, there is thus a likelihood that the level of capital buffer of an Islamic bank is higher than the level of capital buffer of a conventional bank (Maatoug et al., 2019). 2.2 Previous Findings Previous empirical findings on the cyclicality of capital buffer tend to be ambiguous. Huang & Xiong (2015) find a positive co-movement between capital buffer and the business cycle in China, connoting the presence of capital buffer countercyclicality in the Chinese banking industry. This study also find that state-owned bank have a weaker countercyclicality than others banks. By contrast, the study conducted by Lindquist (2004), Ayuso et al. (2004), Coffinet et al. (2012), Marcucci & Quagliariello (2008) and Jokipii & Milne (2008) respectively find a negative co-movement between capital buffer Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 7 and the business cycle in Norwegia, Spain, France, Italy and Europe in general, connoting the presence of capital buffer procyclicality in these countries. Using data from the United States and 29 OECD countries respectively, Shim (2013) and Bikker & Metzemakers (2005) also find that the relationship between capital buffer and the business cycle is negative, implying the existence of capital buffer procyclicality. From Indonesia (Prasetyantoko & Soedarmono, 2010) and Southeast Asian countries in general (Deriantino, 2011), evidence has been presented that the effect on capital buffer of the business cycle is negative and that capital buffer is procyclical. In a study covering sample from 171 developed and developing countries, Chen et al. (2014) report that capital buffer tends to be procyclical in both developed and developing countries. This is true particularly in the cases of bank holding companies and commercial banks. In the case of savings banks and cooperative banks, the effect on capital buffer of the business cycle is subject to the composition of the sample. Meanwhile, Carvallo et al. (2015) based on their study covering 13 Latin American and Caribbean economies report that the cyclical pattern of capital buffer differs across countries. Capital buffer tends to be more procyclical in countries where costs of adjustment are lower and where capital regulation is less stringent. With respect to Islamic banks, empirical findings are also ambiguous. Examining the determinants of Islamic and conventional banks’ capital buffer in the Southeast Asian (ASEAN) and MENA countries over the period 2011- 2015, Maharani & Setiyono (2018) conclude that the relationship between capital buffer and the business cycle is negative. Islamic and conventional banks’ capital buffer is, thus, procyclical. By contrast, in a study that examines the relationship between capital buffer and the business cycle in the MENA countries over the period 2000–2014, it has been found that Islamic and conventional banks’ capital buffers are both countercyclical (Maatoug et al., 2019). It has also been found that the speed of adjustment cost is low for Islamic banks and high for conventional banks. III. Data and Research Methods 3.1 Data. This paper uses secondary panel data that are derived mostly from the end- of-year financial reports of conventional and Islamic banks in Indonesia and from regular reports published by Badan Pusat Statistik (BPS – Statistics Indonesia). The data set is unbalanced and spreads over the period 2004-2019. Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 8 The sample in this study consists of a total 108 banks, of which 11 are Islamic banks and 97 are conventional banks. The sample is selected based on the availability of data. Banks that ceased to exist, or experienced recent conversion, or underwent a recent merger from conventional to Islamic status before the end of the period of analysis are left out of the sample. 3.2 Regression Model The analysis in this paper is conducted using dynamic panel regressions. The regression equations in this paper are provided by Model 1: 𝐶𝐴𝑃𝐵𝑖𝑡 = 𝛽0𝐶𝐴𝑃𝐵𝑖𝑡−1 + 𝛽1𝐶𝑌𝐶𝐿𝐸𝑡 + 𝛽2𝑆𝐼𝑍𝐸𝑖𝑡 + 𝛽3𝐼𝐷𝐼𝑉𝑖𝑡 + 𝛽4𝑃𝑅𝑂𝐹𝑖𝑡 + 𝛽5𝑅𝐼𝑆𝐾𝑖𝑡 + 𝛽6𝑂𝑊𝑁𝐸𝑅𝑖𝑡 + 𝛽7𝑇𝐼𝑀𝐸𝑡 + 𝜀𝑖𝑡 (3) Model 2: 𝐶𝐴𝑃𝐵𝑖𝑡 = 𝛽0𝐶𝐴𝑃𝐵𝑖𝑡−1 + 𝛽1𝐶𝑌𝐶𝐿𝐸𝑡 + 𝛽2𝑆𝐼𝑍𝐸𝑖𝑡 + 𝛽3𝐼𝐷𝐼𝑉𝑖𝑡 + 𝛽4𝑃𝑅𝑂𝐹𝑖𝑡 + 𝛽5𝑅𝐼𝑆𝐾𝑖𝑡 + 𝛽6𝑂𝑊𝑁𝐸𝑅𝑖𝑡 + 𝛽7𝑇𝐼𝑀𝐸𝑡 + 𝛽8𝐼𝑆𝐿𝐴𝑖 + 𝜀𝑖𝑡 (4) Model 3: 𝐶𝐴𝑃𝐵𝑖𝑡 = 𝛽0𝐶𝐴𝑃𝐵𝑖𝑡−1 + 𝛽1𝐶𝑌𝐶𝐿𝐸𝑡 + 𝛽2𝑆𝐼𝑍𝐸𝑖𝑡 + 𝛽3𝐼𝐷𝐼𝑉𝑖𝑡 + 𝛽4𝑃𝑅𝑂𝐹𝑖𝑡 + 𝛽5𝑅𝐼𝑆𝐾𝑖𝑡 + 𝛽6𝑂𝑊𝑁𝐸𝑅𝑖𝑡 + 𝛽7𝑇𝐼𝑀𝐸𝑡 + 𝛽8𝐼𝑆𝐿𝐴𝑖 + 𝛽9𝐼𝑁𝑇𝐸𝑅𝑖𝑡 + 𝜀𝑖𝑡 (5) where 𝐶𝐴𝑃𝐵𝑖𝑡 denotes the capital buffer of bank 𝑖 at time 𝑡, 𝐶𝑌𝐶𝐿𝐸𝑡 denotes the business cycle in the economy, 𝐼𝑆𝐿𝐴 denotes a dummy variable for Islamic banks and 𝐼𝑁𝑇𝐸𝑅 denotes an interaction term between the business cycle and Islamic bank dummy. The dependent variable, capital buffer, is measured as the difference between a bank’s capital adequacy ratio (CAR) and the minimum capital requirement that is imposed on the bank according to its risk profile. Business cycle as the key independent variable is measured as the percentage growth of real gross domestic product (GDP). The other key independent variable, i.e. the dummy for Islamic banks, is given the value 1 for banks that operate according to the interpretation of Islamic laws and 0 otherwise. 𝑆𝐼𝑍𝐸 denotes bank size, 𝐼𝐷𝐼𝑉 denotes income diversification, 𝑃𝑅𝑂𝐹 denotes profitability and 𝑅𝐼𝑆𝐾 denotes credit risk. 𝑂𝑊𝑁𝐸𝑅 denotes a vector of ownership dummy variables, 𝑇𝐼𝑀𝐸 denotes a time variable dummy, 𝛽 denotes the parameters to be estimated and 𝜀 denotes the error term. Bank size, credit risk, profitability and income diversification serve as bank specific control variables. Bank size, which is measured as the natural Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 9 logarithm of total assets, is expected to have a negative sign based on the “too big to be fail” theory. This theory suggests that larger banks dare to take greater risks and are less worry about their financial condition, so that they tend to have lower level of capital buffer (Chen et al., 2014). Credit risk, which is measured as the ratio of non-performing loans to total loans, is expected to have a positive effect on capital buffer. Higher credit risk increases the probability of failure and, hence, increases the level of capital buffer (Maharani & Setiyono, 2018). Profitability, which is measured as the return on equity, is expected to have a positive relationship to capital buffer. In order to meet the minimum capital requirements, banks prefer to use retained earnings rather than external finance (Shim, 2013). Income diversification, which is measured as Herfindahl-Hirschman Index (HHI) of bank’s operating income, is expected to have a negative coefficient. Income diversification reduces the volatility of bank profits and results in less risk. This allows a bank to maintain lower level of capital buffer (Chen et al., 2014). The vector of ownership dummy variables serves as other bank-specific control variables. It comprises a dummy for banks that are belong to the central government, a dummy for banks that are belong to provincial governments and a dummy for foreign banks. Further, the vector of ownership dummy variables also comprises a dummy for ownership concentration, which is given the value 1 for banks whose largest shareholder owns 50 percent or more of shareholdings. Ownership has been reported to have a significant effect on capital buffer (Jiang et al., 2019; Klein et al., 2021). By controlling for these variables, concerns about omitted variable bias can be minimized. The time dummy variable is included to control for time-specific fixed effects. Instead of dummies for years which suffer from collinearity issues, the dummy included is one that represents the period after the Basel III framework is effectively implemented. Its value 1 for the years 2015 onward and 0 otherwise. 3.3 Estimation Method To implement the dynamic panel regressions, this paper employs the popular system generalized method of moments (system GMM) estimators (Arellano & Bover, 1995; Blundell & Bond, 1998). These estimators account for potential endogeneity that takes place when there is a correlation between the independent variables and error terms in the model. These estimators also account for any remaining omitted variable bias and unobserved panel heterogeneity. The business cycle, the dummy for Islamic banks and the interaction term between the business cycle and the Islamic bank dummy are regarded as Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 10 strictly exogenous in the regressions. Similarly, all dummy control variables are regarded as strictly exogenous. Bank-specific control variables that are non- dummies are regarded as predetermined. The system GMM estimators are valid only if the instruments included in the regressions are valid and not correlated with the residuals. To ensure that such condition is met, this paper applies the standard test of overidentifying restrictions developed by Hansen (1982). Besides, to ensure that no second- order serial correlation exists in the error terms, this paper applies the Arellano-Bond tests for zero autocorrelation. IV. Resultsaand Analysis) 4.1 Results.i Table 1 presents the descriptive statistics of the sample. The level of capital buffer in the sample ranges from -2.98 to 155.31 percent, with a mean that is equal to 15.91 percent. The level of capital buffer of Islamic banks is less diverged than the level of capital buffer of conventional banks. The average level of capital buffer of Islamic banks is also lower than that of the level of capital buffer of conventional banks. Table 2 presents the coefficients of correlation between independent variables. For brevity reason, only the correlation coefficients between business cycle and non-dummy bank-specific control variables are displayed. Overall, the correlation coefficients are relatively small, confirming the absence of any multicollinearity issue. Table 3 reports the regression results. In column 1, the regression includes lagged dependent variable, business cycle as the key independent variable, bank size, income diversification, profitability, and credit risk. In column 2, a dummy for Islamic banks is added into the regression. In column 3, an interaction term between the business cycle and the Islamic bank dummy is also added into the regression. The p-values presented at the bottom of the Table indicate that the Hansen test failed to reject the null hypothesis of over- identifying restrictions. Similarly, the Arellano-Bond test for zero autocorrelation fails to reject the null hypothesis of the nonexistence of a second-order serial correlation in the error terms. It is thus confirmed that the system GMM estimators being employed are valid. Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 11 Table 1. Descriptive Statistics N Obs. Mean St. Dev. Min. Max Bank-specific Variables: Con. and Islamic Banks Capital buffer 1543 15.913 16.848 -2.980 155.310 Bank size 1543 15.942 1.798 10.953 21.018 Income diversification 1543 0.187 0.127 0.004 0.500 Profitability 1543 2.023 2.451 -15.820 41.900 Credit risk 1543 2.941 3.053 0.000 50.960 Bank-specific Variables: Conventional Banks Only Capital buffer 1431 16.411 17.252 -2.980 155.310 Bank size 1431 15.929 1.837 10.953 21.018 Income diversification 1431 0.187 0.129 0.006 0.500 Profitability 1431 2.087 2.415 -15.820 41.900 Credit risk 1431 2.872 3.042 0.000 50.960 Bank-specific Variables: Islamic Banks Only Capital buffer 112 9.549 8.071 0.300 53.980 Bank size 112 16.101 1.198 12.901 18.537 Income diversification 112 0.182 0.100 0.004 0.468 Profitability 112 1.200 2.749 -10.770 13.580 Credit risk 112 3.828 3.060 0.100 22.040 Business Cycle and Other Variables Business cycle 1543 5.445 0.527 4.630 6.350 Islamic bank dummy 1543 - - 0 1 D. Ownership types 1543 - - 0 1 D. Ownership concentration 1543 - - 0 1 D. Global financial crisis 1543 - - 0 1 D. Basel implementation 1543 - - 0 1 Table 2. Correlation Matrix Business c. Bank size Income div. Profitability Credit risk Business cycle 1.000 Bank size -0.160 1.000 Income div. -0.021 0.425 1.000 Profitability 0.109 0.072 0.022 1.000 Credit risk -0.061 -0.002 0.050 -0.323 1.000 The coefficient of the business cycle is always negative in columns 1-3 and statistically significant at the 10 percent level. In column 2, the coefficient of the dummy for Islamic banks is not statistically significant. In column 3, the coefficients of the dummy for Islamic banks and its interaction term with the business cycle are both not statistically significant. Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 12 Table 3. Results from the Regressions Model 1 Model 2 Model 3 (1) (2) (3) Lagged capital buffer 0.738** 0.738** 0.738** (0.284) (0.297) (0.297) Business cycle -1.411* -1.392* -1.376* (0.763) (0.771) (0.778) Bank size -4.967** -5.025** -5.028** (2.076) (2.108) (2.107) Income diversification 32.428* 31.836* 31.940* (16.647) (16.428) (16.538) Profitability -2.651 -2.705 -2.682 (2.499) (2.594) (2.603) Credit risk -0.455 -0.490 -0.503 (0.971) (0.964) (0.974) Islamic bank dummy -0.599 0.804 (2.675) (13.753) Business cycle*Islamic bank d. -0.255 (2.313) Ownership type dummies Yes Yes Yes D. Global financial crisis Yes Yes Yes D. Basel III framework Yes Yes Yes Constant 88.401** 89.549** 89.474** (37.476) (38.093) (38.138) N Observations 1,543 1,543 1,543 N Banks 108 108 108 N instruments 17 18 19 Hansen test (p-value) 0.644 0.640 0.645 AR(2) test (p-value) 0.570 0.575 0.576 F-stats (p-value) 0.000 0.000 0.000 Note: The dependent variable is capital buffer. The values reported for each variable are coefficients and heteroscedasticity-autocorrelation-robust standard errors. *, ** and *** indicates significance at the 10, 5 and 1 percent level. 4.2 Robustness Test The regressions reported in Table 3 do not take into account the difference between pre- and post-Basel III economic and regulatory environments. This may bias the results particularly if the effect of this difference correlates with one or more of the existing independent variables. To cope with such a concern, a dummy for Basel III framework is added into the regressions. Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 13 Table 4. Results from the Regressions Model 1 Model 2 Model 3 Lagged capital buffer 0.671*** 0.669*** 0.670*** (0.222) (0.222) (0.223) Business cycle -1.402* -1.398* -1.371* (0.731) (0.732) (0.751) Bank size -4.727* -4.765* -4.780* (2.621) (2.680) (2.690) Income diversification 30.630* 30.427* 30.598* (15.839) (15.565) (15.693) Profitability -2.256 -2.260 -2.250 (1.904) (1.955) (1.954) Credit risk -0.752 -0.784 -0.794 (0.826) (0.802) (0.808) Level of intermediation -4.411 -4.796 -4.584 (21.259) (21.405) (21.538) Islamic bank dummy -0.083 2.301 (1.885) (10.901) Business cycle*Islamic bank d. -0.442 (1.965) Ownership type dummies Yes Yes Yes D. Ownership concentration Yes Yes Yes D. Global financial crisis Yes Yes Yes D. Basel III framework Yes Yes Yes Constant 89.391** 90.317** 90.233** (35.477) (36.406) (36.357) N Observations 1,543 1,543 1,543 N Banks 108 108 108 N instruments 20 21 22 Hansen test (p-value) 0.744 0.743 0.746 AR(2) test (p-value) 0.569 0.571 0.571 F-stats (p-value) 0.000 0.000 0.000 Note: The dependent variable is capital buffer. The values reported for each variable are coefficients and heteroscedasticity-autocorrelation-robust standard errors. *, ** and *** indicates significance at the 10, 5 and 1 percent level. Table 4 reports the results. The coefficient of the business cycle remains negative and significant at the 10 percent level. The coefficient of the dummy for Islamic banks remains not statistically significant and so is the coefficient of the interaction term between the business cycle and the dummy for Islamic banks. Preliminary check indicates the presence of some outliers in the upper side of the dependent variable. Therefore, in addition to the regressions using original data, regressions involving an upper 99th percentile winsorization technique are implemented as a robustness test. Evidence from these regressions suggests that the results above are robust. Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 14 4.3 Discussions The results above consistently show that the relationship between capital buffer and the business cycle in Indonesia is significant negative. The capital buffer of Islamic and conventional banks falls during economic upturns and rises during economic downturns. There is thus evidence for the presence of capital buffer procyclicality in this country. The results above are in line with the findings by Prasetyantoko & Soedarmono (2010) and Deriantino (2011). The results above are also in line with the previous findings by Lindquist (2004), Ayuso et al. (2004), Bikker & Metzemakers (2005), Marcucci & Quagliariello (2008), Jokipii & Milne (2008), Coffinet et al. (2012), Shim (2013) that there is a negative co-movement between capital buffer and the business cycle. The results above further lend a support to the findings by Chen et al. (2014) that capital buffer tends to be more procyclical in developed and developing countries regardless of their level of economic development. The results reported in Table 3 and 4 suggest that capital buffer in Indonesia –a developing country– is procyclical. In relation to Islamic banks, the results above are in agreement with the findings by Maharani & Setiyono (2018) and against the ones by Maatoug et al. (2019). The fact that none of the coefficients of the dummy for Islamic banks and of the interaction term between this variable and the business cycle are significant implies that no difference exists in the level of capital buffer of Islamic banks as compared to conventional banks. Similarly, the fact that none of the coefficients of the dummy for Islamic banks and of the interaction term between this variable and the business cycle are significant implies that no difference exists in the cyclicality of capital buffer in Islamic and conventional banking. Thus, despite their uniqueness for not paying or charging interest and instead performing intermediation function using Islamic financial contracts (Askari et al., 2015; Habib, 2018), Islamic banks have a relatively similar capital buffer behavior to conventional banks. This echoes some previous findings that there tends to be no significant difference in the behavior of Islamic and conventional banks. For example, in term of bank margin (Susamto et al., 2021), efficiency (El-Gamal & Inanoglu, 2005; Hassan et al., 2009) and stability (Kasri & Azzahra, 2020). V. Concluding Remarks 5.1 Conclusions This paper has analyzed the cyclical behavior of capital buffer of Islamic and conventional banks in Indonesia using dynamic panel regressions. In line with Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 15 theoretical predictions, the results provide evidence that the capital buffer of Indonesian banking industry is significantly affected by the business cycle. The capital buffer of Indonesian Islamic and conventional banks as a whole is procyclical. The results however fail to provide evidence for the presence of a difference in the level of capital buffer of Islamic banks as compared to conventional banks. The results also fail to provide evidence that there is a difference in the cyclicality of capital buffer of Islamic and conventional banks. Thus, different from theoretical predictions that the level of capital buffer of Islamic banks is higher than the level of capital buffer of conventional banks, Islamic banks have a relatively similar capital buffer behavior to conventional banks. 5.2 Recommendation The results in this paper suggest a deviation from the Basel III regulations. Therefore, if the countercyclical capital buffer as determined by the Basel III framework is to be achieved, a policy measure aimed to alter the cyclical behavior of capital buffer of Islamic and conventional banks in Indonesia is a must. The results in this paper also suggest that such policy measure needs not to be specified for Islamic or conventional banking industry. Aside from Islamic banks’ distinct characteristics, the level and the cyclicality of capital buffer in Islamic banking industry are not significantly different from the level and the cyclicality of capital buffer in conventional banking industry. This paper relies on annual growth of real GDP to measure business cycle. This may not fully capture the effect of the business cycle. Future research should attempt to examine the cyclical behavior of capital buffer using more frequent dataset. Future research should also attempt to look deeper into the details of the cyclical behavior of capital buffer before and after the Basel III framework is implemented. References Acharya, S. (1996). Charter value, minimum bank capital requirements and deposit insurance pricing in equilib- rium. Journal Banking Finance, 20, 351–375. Amato, J. D., & Furfine, C. H. (2004). Are credit ratings procyclical? Journal of Banking and Finance, 28(11), 2641–2677. https://doi.org/10.1016/j.jbankfin.2004.06.005 Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 16 Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of Econometrics. https://doi.org/10.1016/0304-4076(94)01642-D Askari, H., Iqbal, Z., & Mirakhor, A. (2015). Introduction to Islamic Economics- Theoy anf Application. John Wiley & Sons Singapore Pte. Ltd. Ayuso, J., Pérez, D., & Saurina, J. (2004). Are capital buffers pro-cyclical? Evidence from Spanish panel data. Journal of Financial Intermediation, 13(2), 249–264. https://doi.org/10.1016/S1042-9573(03)00044-5 BCSB. (2010). Basel III: A Global Regulatory Framework for More Resilient Banks and Banking Systems. Bank for International Settlement (BIS). Bikker, J. A., & Metzemakers, P. A. J. (2005). Bank provisioning behaviour and procyclicality. Journal of International Financial Markets, Institutions and Money, 15(2), 141–157. https://doi.org/10.1016/j.intfin.2004.03.004 Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87(1), 115–143. https://doi.org/10.1016/S0304-4076(98)00009-8 Bougatef, K., & Korbi, F. (2018). Capital buffer and credit-risk adjustments in Islamic and conventional banks. Thunderbird International Business Review, 1–18. https://doi.org/10.1002/tie.22022 Carvallo, O., Kasman, A., & Kontbay-Busun, S. (2015). The Latin American bank capital buffers and business cycle: Are they pro-cyclical? Journal of International Financial Markets, Institutions and Money, 36(C), 148– 160. Chen, Y.-K., Ho, A. Y.-F., & Hsu, C.-L. (2014). Are bank capital buffers cyclical? Evidence of developed and developing countries. Journal of Financial Studies, 22(3), 27–55. Coffinet, J., Coudert, V., Pop, A., & Pouvelle, C. (2012). Two-way interplays between capital buffers and credit growth: Evidence from French banks. Journal of International Financial Markets, Institutions and Money, 22(5), 1110–1125. https://doi.org/10.1016/j.intfin.2012.05.011 Daher, H., Masih, M., & Ibrahim, M. (2015). The unique risk exposures of Islamic banks’ capital buffers: A dynamic panel data analysis. Journal of International Financial Markets, Institutions and Money, 36, 36–52. https://doi.org/10.1016/j.intfin.2015.02.012 de Carvalho, F. A., & de Castro, M. R. (2015). Macroprudential and Monetary Policy Interaction: a Brazilian perspective. In Banco Central do Brasil. De Mooij, R. A. (2011). Tax Biases to Debt Finance: Assessing the Problem, Finding Solutions. Fiscal Studies, 33(4), 489–512. https://doi.org/10.1111/j.1475-5890.2012.00170.x Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 17 Deriantino, E. (2011). Procyclicality of banks’ capital buffer and its impact on bank’s lending activity in ASEAN countries. In Working Paper. https://doi.org/10.2139/ssrn.1894488 El-Gamal, M. A., & Inanoglu, H. (2005). Inefficiency and heterogeneity in Turkish banking: 1990–2000. Journal of Applied Econometrics, 20(5), 641–664. https://doi.org/https://doi.org/10.1002/jae.835 Guidara, A., Lai, V. S., Soumaré, I., & Tchana, F. T. (2013). Banks’ capital buffer, risk and performance in the Canadian banking system: Impact of business cycles and regulatory changes. Journal of Banking and Finance, 37(9), 3373–3387. https://doi.org/10.1016/j.jbankfin.2013.05.012 Habib, S. F. (2018). Fundamentals of Islamic Finance and Banking. John Wiley & Sons. Hansen, L. P. (1982). Large sample properties of generalized method of moments estimators. Econometrica, 50(4), 1029–1054. https://doi.org/doi:10.2307/1912775 Hassan, T., Mohamad, S., & Khaled I. Bader, M. (2009). Efficiency of conventional versus Islamic banks: evidence from the Middle East. International Journal of Islamic and Middle Eastern Finance and Management, 2(1), 46–65. https://doi.org/10.1108/17538390910946267 Huang, X., & Xiong, Q. (2015). Bank capital buffer decisions under macroeconomic fluctuations: Evidence for the banking industry of China. International Review of Economics and Finance, 36(12), 30–39. https://doi.org/10.1016/j.iref.2014.11.005 Jiang, C., Liu, H., & Molyneux, P. (2019). Do different forms of government ownership matter for bank capital behavior? Evidence from China. Journal of Financial Stability, 40, 38–49. https://doi.org/https://doi.org/10.1016/j.jfs.2018.11.005 Jokipii, T., & Milne, A. (2008). The cyclical behaviour of European bank capital buffers. Journal of Banking and Finance, 32(8), 1440–1451. https://doi.org/10.1016/j.jbankfin.2007.12.001 Kane, E. J. (1989). Changing incentives facing financial-services regulators. Journal of Financial Services Research, 2(3), 265–274. https://doi.org/10.1007/BF00130743 Kasri, R. A., & Azzahra, C. (2020). Do Islamic banks more stable than conventional banks? Evidence from Indonesia. Jurnal Ekonomi & Keuangan Islam, 6(2 SE-Articles), 149–164. https://doi.org/10.20885/jeki.vol6.iss2.art6 Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 18 Khan, F. (2010). How “Islamic” is Islamic Banking? Journal of Economic Behavior and Organization, 76(3), 805–820. https://doi.org/10.1016/j.jebo.2010.09.015 Klein, P., Maidl, C., & Woyand, C. (2021). Bank ownership and capital buffers: How internal control is affected by external governance. Journal of Financial Stability, 54, 100857. https://doi.org/https://doi.org/10.1016/j.jfs.2021.100857 Lindquist, K. G. (2004). Banks’ buffer capital: How important is risk. Journal of International Money and Finance, 23(3), 493–513. https://doi.org/10.1016/j.jimonfin.2004.01.006 Maatoug, A. Ben, Ayed, B. W., & Zied, F. (2019). Are MENA banks’ capital buffers countercyclical? Evidence from the Islamic and conventional banking systems. Quarterly Review of Economics and Finance, 74, 109–118. https://doi.org/10.1016/j.qref.2019.04.006 Maharani, N. K., & Setiyono, B. (2018). Do Risk, Business Cycle, and Competition Affect Capital Buffer? an Empirical Study on Islamic Banking in Asean and Mena. In Journal of Islamic Monetary Economics and Finance (Vol. 3, Issue 2). https://doi.org/10.21098/jimf.v3i2.888 Marcucci, J., & Quagliariello, M. (2008). Is bank portfolio riskiness procyclical? Evidence from Italy using a vector autoregression. Journal of International Financial Markets, Institutions and Money, 18(1), 46–63. https://doi.org/10.1016/j.intfin.2006.05.002 Mcnally, W. J. (1999). Market Open Signaling Repurchase. Financial Managementment, 28(2), 55–67. Milne, A. K. L. (2004). The Inventory Perspective on Bank Capital. SSRN Electronic Journal, August, 1–43. https://doi.org/10.2139/ssrn.576062 Milne, A., & Whalley, A. E. (2011). Bank Capital Regulation and Incentives for Risk-Taking. SSRN Electronic Journal, December. https://doi.org/10.2139/ssrn.299319 Myers, S. ., & Majluf, N. (1984). Corporate financing and investment decisions when firms have information investors do not have. Journal of Financial Economics 13, 187–221. Nier, E., & Baumann, U. (2006). Market discipline, disclosure and moral hazard in banking. Journal of Financial Intermediation 15, 332–361. P Jackson, Furfine, C., C. Groeneveld, D. Jones, W. Perraudin, L. Radecki, & M. Yoneyama. (1999). Capital Requirements and Bank Behaviour: The Impact of the Basel Accord. Basel Committee on Banking Supervision,. Working Paper No. 1. Prasetyantoko, A., & Soedarmono, W. (2010). The determinants of capital buffer in Indonesian banking. Financial Stability Review, Forthcomin. Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 19 Shim, J. (2013). Bank capital buffer and portfolio risk: The influence of business cycle and revenue diversification. Journal of Banking and Finance, 37(3), 761–772. https://doi.org/10.1016/j.jbankfin.2012.10.002 Stolz, S., & Wedow, M. (2011). Banks’ regulatory capital buffer and the business cycle: Evidence for Germany. Journal of Financial Stability, 7(2), 98–110. https://doi.org/10.1016/j.jfs.2009.09.001 Susamto, A. A., Octavio, D. Q., & Wardani, D. T. K. (2021). Benchmarking the intermediation costs of Islamic and conventional banks : Evidence from Indonesia. Journal of Economic Cooperation and Development, 42(3), 91–118. Winter, R. A. (1994). The Dynamics of Competitive Insurance Markets. In Journal of Financial Intermediation (Vol. 3, Issue 4, pp. 379–415). https://doi.org/10.1006/jfin.1994.1011 Atikah, Susamto, Saleh, & Abdic | The cyclical behavior of capital buffer of Indonesian Islamic and conventional banks Internasional Journal of Islamic Economics and Finance (IJIEF), 4(1), 1-20│ 20 This page is intentionally left blank