This study investigates if the implementation of International Financial Reporting Standard (IFRS)-9: Financial Instruments (IFRS 9 hereafter) affects a firm’s cash holdings from a developing country perspective. Moreover, we explore whether the above nexus varies between Islamic banks and conventional banks in the same setting.
This study covers all the listed banks in Bangladesh. The data period is 2015–2022, which allows the study to have a pre- and post-IFRS 9 impact on corporate cash holdings. We use ordinary least square regression models to test our conjectures. Our entire analysis is based on 232 firm-year observations.
The overall findings suggest that the cash holding decreased significantly in post-IFRS 9 periods compared to pre-IFRS 9 periods. We further test whether the impact of IFRS 9 presents heterogeneity between Islamic banks and conventional banks in terms of cash holdings. However, we do not find any variation. Our results remain robust through a set of alternative measures of cash holding and sub-sample analysis.
Our study presents an empirical analysis of IFRS 9 in general, and in a developing country Bangladesh in particular. Prior research overlooked the possible impact of IFRS 9 from a developing country perspective, hence, this paper contributes to policy development and the literature of IFRS in emerging countries.
1. Introduction
During the period of the global financial crisis of 2008, the impairment calculation measures of International Accounting Standards (IAS) came under criticism. It was argued that these measures caused credit loss recognition delays, leading to insufficient allowance. Regulators paid attention to this criticism since credit risks were considered a significant factor behind the crisis (e.g. Bhatt et al., 2023; Shin and Kim, 2015). In response to this criticism, in July 2014, the International Accounting Standards Board (IASB) issued the final version of IFRS 9 (IASB, 2011), which includes the new impairment rules. The IAS 39 and the IFRS 9 impairment models in principle only recognize credit losses and thus exclude losses from other risk factors and are restricted to expected losses and thus do not provide for the possibility of unexpected credit losses (Gornjak, 2020; Karpuz et al., 2020; Albrahimi, 2019; Gebhardt, 2016). The implementation of this new impairment model is a fundamental shift from the way companies used to treat credit risks. Instead of incorporating losses as they occur, it requires firms to estimate and record expected credit losses due to credit loss or default for all instruments held at amortized cost or fair value through other comprehensive income (Kvaal et al., 2023). The new IFRS 9 impairment rules will thus affect available capital, and alignment of data and approach between the firms' strategies and policies regarding credit risk, finance and working capital management, to be specific (Giner and Mora, 2021; Albrahimi, 2019). The influence of IFRS 9 on cash holdings thus deserves research attention.
In accounting research, there are two contrasting viewpoints regarding the impact of IFRS on corporate cash holdings (see Akgün and Karataş, 2021; Morshed, 2020). The first viewpoint demonstrates that IFRS adoption increases firms’ transparency, disclosure and creditworthiness, which helps in accessing external capital and credit. As a result, firms need to hold less cash (Ozkan et al., 2021). On the other hand, by taking into account the revolutionary forward-looking impairment recognition of IFRS 9, the second viewpoint argument demonstrates that cash holding might also increase in firms due to the restrictions on credit sales and encouragement of buying on a cash basis (Ozkan et al., 2021).
Previous research has suggested that the advantages of implementing IFRS may be influenced by various factors such as cultural disparities, the effectiveness of legal and political systems in enforcing accounting and auditing standards and the existence of strong incentives for companies to provide reliable and accurate financial reports. These findings have further propelled this study to examine the impact of IFRS 9 in emerging markets, where the corporate environments significantly differ from the markets in developed countries that have received the most attention from researchers (see Tlemsani et al., 2024; Durán et al., 2016; Al Sawalqa and Qtish, 2021; Damak-Ayadi et al., 2020; Tawiah and Boolaky, 2020; Daske et al., 2008).
Our research paper adds to the growing body of literature that examines the impact of IFRS on various aspects of accounting and financial management systems. We contribute to this literature by including IFRS 9 (instead of “IFRS adoption” as the independent variable) in our models and analyzing the effect of the new impairment model on cash holding patterns among financial institutions in an emerging market. We focus on Bangladesh as a case study to investigate the effect of IFRS 9, which will provide new insights into the existing literature on the economic consequences of IFRS 9. For instance, this study is the first attempt to examine the impact of IFRS 9 on cash holdings in a country setting that is different from the existing research in the literature. Bangladesh is a country with high growth potential, but it is still in the early stages of economic development (Chen et al., 2023). The institutional framework incorporated the financial reporting regime, but it is not yet fully prepared to ensure proper disclosure and transparency. Hence, the existing empirical evidence is likely to differ from other countries. Moreover, the majority of the literature on this topic focuses on firms in developed countries, where firms have a strong relationship with the capital market and are supported by a strong legal and regulatory system (i.e. Manoel et al., 2018; Steijvers and Niskanen, 2013). In contrast, access to capital, credit or trade credit, facilitating trade credit, is much more challenging in countries like Bangladesh due to the absence of an adequate institutional framework. The institutional environment is facing several challenges due to a weak regulatory framework (Ali and Ahmed, 2007; Nurunnabi, 2017; Ferdous et al., 2014). Additionally, multiple capital market collapses have occurred, further eroding the trust of stakeholders in firms. Therefore, it is important to study the cash-holding behavior and the impact of IFRS 9 as a separate case study.
Secondly, our research differs from most studies in terms of ownership structure. The majority of studies focused on publicly listed companies with diffused ownership, which is not the case in countries like Bangladesh, where family-owned companies dominate the corporate sector (Badrul Muttakin et al., 2014). Shareholders in such companies are often concentrated in a single family (Miah et al., 2023), leading to less transparency and higher levels of information asymmetry. As a result, the amount of cash to be held often becomes a controversial issue, exacerbating agency problems due to the existence of market imperfections. Family firms typically have control over working capital management, and decisions related to cash management play a significant role in the firm’s financial policies (Steijvers and Niskanen, 2013; Mahmud et al., 2022). Thirdly, Islamic banks in Bangladesh must be Shariah compliant and the banks are run by an Independent Shariah Supervisory Committee (Bangladesh Bank, 2009). However, recently, it is found that sponsor shareholders or members of the board of directors are accused of severe money laundering and fund embezzlement in those publicly listed Islamic banks [1]. Hence, it is timely to investigate whether regulatory initiative by accounting standard setters (such as IASB) affects firm-level cash holdings, which is covered in the present study.
Finally, a comprehensive study on the impact of IFRS 9 on the financial liquidity of companies during financial crises is of paramount importance (Gebhardt, 2016; Carvalhal and Leal, 2013). The case of Bangladesh is particularly valuable in this regard, given the country’s recent experiences of institutional collapses, especially in the banking sector and the delisting of companies by the Bangladesh Securities and Exchange Commission (BSEC). The number of financially distressed firms in Bangladesh seems to be increasing over time. Furthermore, the country’s financial institutions are grappling with significant bad debts (Islam and Ullah, 2020), which, coupled with the recent crisis in foreign reserves, have threatened their creditworthiness. These developments have put immense pressure on regulatory bodies to focus on impairment issues and overall financial reporting in Bangladesh. Therefore, the findings of this study are expected to contribute significantly to the theoretical and practical understanding of the consequences of IFRS 9 for financial institutions. In particular, standard setters, regulators and policymakers can leverage the insights generated by this study to facilitate the implementation and revision of the standard and ensure better compliance across the industry. More specifically, the Bangladesh bank can utilize our paper’s findings to formulate policies for conventional banks and for Islamic banks.
The overall findings provide evidence that the implementation of IFRS is associated with a decrease in cash holding, which is consistent with the prior studies (Ozkan et al., 2021; Farinha et al., 2018). The results suggest that the financial institutions of Bangladesh have implemented IFRS 9 and amended their working capital policies. These changes are most likely to have mitigated the large volatility in earnings caused by large bad debt provisions. Second, our result is attributed to the improvement of the information environment due to IFRS adoption (Daske et al., 2008), and the reduction of earnings manipulations or improvement of accounting earnings quality (Ahmed et al., 2013; Yi Lin et al., 2012). Given the rising trend of Islamic banking practices in developing countries, we attempt to isolate the impact of IFRS 9 on cash holdings on Islamic banks over conventional banks. Our results remain robust regardless of the types of banks. Hence, we can infer that the negative impact of IFRS 9 on cash holding does not vary whether a bank belongs to the conventional banking system or a bank belongs to the Shariah-based banking system. Our results remain valid in a set of sensitivity analyses.
The paper is structured as follows. In section 2, the hypothesis is developed based on the argument of why IFRS is likely to increase or decrease cash holdings. The research design is presented in the next Section 3 followed by the research findings and sensitivity analyses in Section 4. Section 5 contains further analysis to understand the impact of bank size, loans and investment and age on the bank's cash holdings. Finally, the conclusion, policy implications and suggestions for future research are presented in Section 6.
2. Literature review and hypothesis development
2.1 Review of previous literature on IFRS 9 and corporate cash holdings
IFRS 9 introduces a radical change in recognizing impairment on financial instruments, including trade receivables. This radical change in impairment calculation has caused two contrasting effects on the cash holdings of firms. On one hand, the empirical evidence explains that the adoption of IFRS 9 improves accounting quality, which leads to a decrease in the amount of cash held by the firms (Ozkan and Ozkan, 2004; Gebhardt, 2016; Giner and Mora, 2019). Using the propositions of the agency theory, this literature suggests that the early recognition of impairment (based on expected credit loss increases transparency and earnings quality, which in turn reduces information asymmetry between managers and shareholders, thus self-interested agents/managers have reduced scope to hold unnecessary cash. The new impairment calculation also requires information relating to the creditworthiness of the firm, which increases the accounting quality of firms, which leads to better transparency of the financial reports of the firms, thereby reducing information asymmetry and as a result, the creditworthiness of the firms increases. Moreover, the expected credit loss model also requires firms to collect information on corporate governance mechanisms. Thus, well-governed firms can reduce the information asymmetry between capital investors and agents (Ozkan and Ozkan, 2004; Opler et al., 1999). It is argued that accounting quality helps firms to reduce agency problems, ensure transparency, and, in turn, is perceived to ensure effective monitoring of the agents to prevent firms from holding more cash to serve their interests. Overall, this increased transparency and earnings quality lead to high creditworthiness, which lowers to cost of external funds. Post-implementation of IFRS 9, firms are thus supposed to have better access to capital and credit, which reduces the firm’s predisposition to hold more cash (Ozkan et al., 2021; Li et al., 2024; Giner and Mora, 2019).
On the contrary, evidence also suggests that the implementation of IFRS 9 changes the credit facilities and sales patterns for firms, thus, increasing in cash holding of firms. The expected credit loss model of IFRS 9 requires firms to recognize sufficient bad debt allowances, which are often significantly higher than what firms used to require under the IAS-39 provisions. This additional allowance naturally causes variation in the retained earnings of the firm. Hence, firms encourage cash sales instead of credit sales or, at best, those trade receivables with a first-track collection provision. As a result, firms' cash holdings increase. Cash holding also increases under these new provisions to those firms that used to buy in credit. To make timely payments under first-track trade payable, or to support cash purchases, firms need to hold more cash post-IFR-9 implementation. Three theories in corporate finance also explain why firms hold more or less cash. The first theory, known as the trade-off theory (TOT), explains that firms hold more cash to avoid transaction costs when raising funds externally or selling assets (Miller and Orr, 1966). Additionally, the theory suggests that managers hold more cash to prepare for unforeseen difficulties that may require cash reserves. The second theory, known as the pecking order theory, explains that firms that rely on internal cash for investments tend to hold more cash than other firms (Chaklader and Padmapriya, 2021; Alnori et al., 2022). Finally, the free cash flow theory, which is based on the agency theory, suggests that managers may hold more cash to serve their interests, rather than the company’s (Ozkan et al., 2018; Smith and Pennathur, 2019). Free cash allows them to pursue their goals, which may not always align with the company’s best interests. In Bangladesh, family-owned companies dominate the corporate sector, and political influences are pervasive among the companies. Unfortunately, the regulatory framework of the country is often criticized for its inadequate ability to address the self-opportunistic motives of agents, which is frequently reflected in poor accounting quality and manipulation of financial reports (Molla et al., 2023; Sobhan, 2021; Sobhan and Bose, 2019).
Among the firm-level determinants of cash holdings such as leverage, firm size, dividend payment, capital expenditure and investment opportunities, have been considered by a good number of studies (Caprio et al., 2020; Magerakis et al., 2020; Ozkan et al., 2021; Dimitropoulos, 2020). Moreover, research interest is also increasing in understanding the effect of accounting quality on the cash holding pattern of firms (Pathak et al., 2022). Regardless of the variables used to measure accounting quality, all of these studies suggest that poor accounting quality leads firms to hold more cash, and the higher the quality, the lesser the amount of cash held by firms (Beatty et al., 2008). The study of Ozkan et al. (2018) adds another dimension to the existing literature by considering the accounting regime effect on the accounting quality-cash holding relationship in Turkey. However, the findings remain the same. The study found a significant and negative relationship between IFRS adoption and cash holdings in post-IFRS adoption periods. Considering the overall positive effects of the IFRS adoption worldwide, and the regulatory and institutional framework of our sample country, which gives a valid justification to test the effect of the IFRS adoption on cash holding patterns among the financial institutions, we developed our first hypothesis, as follows:
The implementation of IFRS 9 is negatively associated with the cash holdings.
2.2 IFRS 9, Islamic banks vs. conventional banks and corporate cash holdings
While our first hypothesis intends to understand the impact of IFRS 9 on cash-holding patterns amongst financial institutions, it also intends to explore if this association of IFRS 9 and cash-holding varies between Islamic banks, which are compliant with the Shariah laws. Islamic corporate finance studies suggest that firms that are Shariah-compliant are subject to several restrictions to keep their Shariah-compliant status (Alnori and Bugshan, 2023; Bugshan et al., 2021; Naz et al., 2017; Yildirim et al., 2018). Most of these restrictions significantly influence a firm’s financing and working capital management strategies. Studies suggest that compare to conventional banks, the Sharia-compliant banks (in the case of Bangladesh, the Islamic banks) find their external financing more costly and limited (Nethercott, 2012). Thus, as the TOT suggests, to avoid the transaction cost of external funds, the Shariah-compliant firms take precautionary motives and hold comparatively more cash. Shariah-compliant firms’ cash management policies are therefore different from conventional firms’ policies. Our study thus expects that in post IFRS 9 periods, cash holdings may exhibit a different effect on the Islamic banks than the conventional banks of Bangladesh. Hence, we developed our second hypothesis as follows:
The impact of IFRS 9 on corporate cash holdings varies between Islamic banks and conventional banks.
3. Research methods
3.1 Sample selection
This study is based on the listed banks in Bangladesh. We start from 2015 to have pre-IFRS 9 impact, and we cover up to 2022 to see the impact of IFRS 9 in post-adoption periods. We use ordinary least square regression models to test our conjectures. Prior research shows that the financial sector is heavily affected by IFRS 9 compared to other sectors in any jurisdiction (Awuye and Taylor, 2024). Hence, our entire analysis is based on the banking sector in Bangladesh. Initially, we started with 33 listed banks in Bangladesh, however, four banks were excluded due to data limitations, leaving us with 29 banks. Our entire analysis is based on 232 firm-year observations. The sample derivation process is explained in Table 1.
3.2 Model specification
To test our hypothesis, we estimate the following regression models following prior research (Francis et al., 2014; Shikimi, 2019).
Where CASH is the only dependent variable, which is measured as the ratio of total cash and marketable securities to total assets of the bank. We assign 1 for IFRS9 for the periods of 2018 and onwards and 0 for the prior periods. We also assign 1 for ISLAMIC if the bank follows a Sharia-based banking system (full-fledged Islamic banks) and 0 for other conventional banks in Bangladesh. All other control variables are defined following prior related literature. Appendix A presents the definitions of all variables used in the regression models.
4. Results and discussion
4.1 Summary statistics
Table 2 presents summary statistics of the full sample of our study. The mean value of CASH for our sample is 0.092. The mean value of IFRS 9 is 0.625, which indicates that more than 62% of firm-year observations come from post-IFRS 9 periods over non-IFRS 9 periods. The mean value of external debt (DEBT) is 0.091. Table 3 provides the Spearman correlation matrix statistics. We find that IFRS 9 has a negative relationship with corporate cash holdings, which is consistent with our hypothesis. In addition, cash holding is a negative deposit ratio, goodwill and is positively associated with firm size, debt level, capital expenditure, capital adequacy, the board size, audit committee size, operating cash flow, financial distress measure and property, plant and equipment etc. Following prior research, we check the Variance Inflation Factor for all the variables, and we confirm that correlations are within the limit, hence, our models do not suffer from multicollinearity problems.
Descriptive statistics
| Variable | N | Mean | P50 | SD | P25 | P75 | P90 | Min | Max |
|---|---|---|---|---|---|---|---|---|---|
| CASH | 232 | 0.092 | 0.084 | 0.037 | 0.066 | 0.113 | 0.140 | 0.000 | 0.222 |
| IFRS 9 | 232 | 0.625 | 1.000 | 0.485 | 0.000 | 1.000 | 1.000 | 0.000 | 1.000 |
| COF | 232 | 0.031 | 0.018 | 0.045 | 0.010 | 0.036 | 0.065 | −0.002 | 0.272 |
| FIRMSIZE | 232 | 19.592 | 19.571 | 0.434 | 19.317 | 19.835 | 20.108 | 18.694 | 21.332 |
| DEBT | 232 | 0.128 | 0.091 | 0.149 | 0.055 | 0.131 | 0.285 | 0.000 | 0.804 |
| CAPEX | 232 | 0.003 | 0.001 | 0.007 | 0.001 | 0.003 | 0.005 | 0.000 | 0.076 |
| LOANS2TA | 232 | 0.432 | 0.615 | 0.330 | 0.000 | 0.708 | 0.742 | 0.000 | 0.814 |
| DEPOSIT2TA | 232 | 0.701 | 0.744 | 0.167 | 0.690 | 0.787 | 0.820 | 0.000 | 0.904 |
| TIER1 | 232 | 0.021 | 0.000 | 0.032 | 0.000 | 0.054 | 0.068 | 0.000 | 0.123 |
| TIER2 | 232 | 0.009 | 0.000 | 0.015 | 0.000 | 0.020 | 0.033 | 0.000 | 0.056 |
| ROA | 232 | 0.007 | 0.007 | 0.004 | 0.005 | 0.010 | 0.012 | −0.004 | 0.018 |
| GOODWILL | 232 | 0.000 | 0.000 | 0.001 | 0.000 | 0.000 | 0.000 | 0.000 | 0.006 |
| SIZE_BOARD | 232 | 14.004 | 14.000 | 4.041 | 11.000 | 17.000 | 20.000 | 6.000 | 24.000 |
| INDBOD | 232 | 2.595 | 2.000 | 1.077 | 2.000 | 3.000 | 4.000 | 0.000 | 6.000 |
| SIZE_AC | 232 | 4.332 | 5.000 | 1.080 | 3.000 | 5.000 | 5.000 | 2.000 | 12.000 |
| INDIRAC | 232 | 1.866 | 2.000 | 0.661 | 2.000 | 2.000 | 3.000 | 0.000 | 3.000 |
| DIVIDEND | 232 | 0.733 | 1.000 | 0.443 | 0.000 | 1.000 | 1.000 | 0.000 | 1.000 |
| OCF | 232 | 0.012 | 0.011 | 0.033 | 0.000 | 0.021 | 0.043 | −0.127 | 0.210 |
| LNAGE | 232 | 3.004 | 2.970 | 0.438 | 2.708 | 3.434 | 3.584 | 2.079 | 3.829 |
| ZSCORE | 231 | 0.224 | 0.203 | 0.174 | 0.130 | 0.257 | 0.338 | −0.003 | 1.026 |
| PPE | 232 | 0.016 | 0.016 | 0.008 | 0.011 | 0.021 | 0.026 | 0.000 | 0.047 |
Source(s): Authors’ own creation
Correlation statistics
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CASH | 1 | 1.000 | |||||||||||||||||||
| IFRS9 | 2 | −0.256 | 1.000 | ||||||||||||||||||
| FIRMSIZE | 3 | 0.015 | 0.511 | 1.000 | |||||||||||||||||
| DEBT | 4 | 0.324 | 0.072 | 0.521 | 1.000 | ||||||||||||||||
| CAPEX | 5 | 0.126 | 0.068 | −0.010 | 0.025 | 1.000 | |||||||||||||||
| LOANS2TA | 6 | 0.026 | −0.026 | −0.188 | −0.103 | −0.094 | 1.000 | ||||||||||||||
| DEPOSIT2TA | 7 | −0.118 | −0.094 | −0.442 | −0.802 | −0.005 | 0.129 | 1.000 | |||||||||||||
| TIER1 | 8 | 0.017 | 0.036 | 0.277 | 0.213 | 0.187 | −0.045 | −0.178 | 1.000 | ||||||||||||
| TIER2 | 9 | 0.108 | 0.074 | 0.255 | 0.199 | 0.273 | −0.108 | −0.163 | 0.840 | 1.000 | |||||||||||
| ROA | 10 | 0.016 | −0.303 | −0.397 | −0.023 | 0.030 | 0.107 | 0.112 | 0.057 | −0.121 | 1.000 | ||||||||||
| GOODWILL | 11 | −0.170 | −0.047 | 0.028 | −0.012 | 0.072 | 0.172 | −0.026 | 0.300 | −0.016 | 0.257 | 1.000 | |||||||||
| SIZE_BOARD | 12 | 0.108 | −0.057 | 0.032 | 0.217 | 0.020 | −0.060 | −0.193 | −0.158 | −0.104 | −0.059 | −0.259 | 1.000 | ||||||||
| INDBOD | 13 | 0.109 | 0.015 | 0.279 | 0.299 | 0.218 | −0.046 | −0.206 | 0.365 | 0.270 | 0.078 | 0.200 | 0.328 | 1.000 | |||||||
| SIZE_AC | 14 | 0.135 | −0.125 | −0.132 | 0.109 | −0.044 | 0.037 | −0.080 | −0.027 | 0.009 | 0.017 | −0.171 | 0.555 | 0.112 | 1.000 | ||||||
| INDIRAC | 15 | 0.052 | −0.022 | −0.062 | 0.033 | 0.022 | −0.036 | −0.008 | 0.144 | 0.175 | −0.004 | −0.070 | 0.185 | 0.319 | 0.293 | 1.000 | |||||
| DIVIDEND | 16 | 0.126 | −0.025 | 0.080 | 0.223 | 0.081 | 0.083 | −0.115 | 0.243 | 0.237 | 0.135 | 0.053 | 0.148 | 0.162 | 0.150 | 0.217 | 1.000 | ||||
| OCF | 17 | 0.164 | −0.138 | −0.144 | 0.052 | 0.409 | 0.122 | −0.032 | 0.197 | 0.193 | 0.068 | 0.047 | −0.132 | 0.105 | 0.012 | 0.132 | 0.051 | 1.000 | |||
| LNAGE | 18 | 0.051 | 0.264 | 0.512 | 0.185 | −0.080 | −0.044 | −0.183 | −0.002 | 0.089 | −0.281 | −0.266 | −0.145 | −0.184 | −0.102 | −0.118 | −0.123 | −0.121 | 1.000 | ||
| ZSCORE | 19 | 0.276 | −0.045 | 0.414 | 0.878 | 0.048 | −0.127 | −0.687 | 0.261 | 0.183 | 0.095 | 0.087 | 0.213 | 0.333 | 0.156 | −0.015 | 0.212 | 0.112 | 0.167 | 1.000 | |
| PPE | 20 | 0.109 | −0.176 | −0.069 | 0.018 | 0.113 | −0.074 | 0.054 | 0.071 | 0.052 | 0.011 | 0.080 | 0.046 | −0.077 | 0.145 | 0.061 | −0.099 | 0.124 | 0.171 | 0.068 | 1.000 |
Source(s): Authors’ own creation
4.2 Main results
Table 4 displays the baseline regression results of IFRS 9 and corporate cash holding. First, show the relationship between IFRS 9 and corporate cash holdings without including control variables. It shows that the coefficient of IFRS 9 is negative and statistically significant at 1% and our results are also economically significant, suggesting that 3.67% of cash holding decreases in post-IFRS 9 adoption periods compared to pre-IFRS 9 periods. The evidence thus supports our conjecture regarding the impact of IFRS 9 on corporate cash holdings. Column 2 shows the results between IFRS 9 and corporate cash holdings with all control variables and Column 3 shows the results of the interaction of IFRS 9 and Islamic banks on corporate cash holdings. Results (in column 2) show that the coefficient of IFRS 9 is negative and statistically significant at 1% which indicates that banks decrease cash holding in post-IFRS 9 adoption periods compared to pre-IFRS 9 periods, which is consistent with the prior research (Farinha et al., 2018; Ozkan et al., 2021). Our result is attributed to the improvement of the information environment due to IFRS adoption (Daske et al., 2008) and the reduction of earnings manipulations (or improvement of accounting earnings quality (Ahmed et al., 2013; Yi Lin et al., 2012). Next, we find that the coefficient of IFRS 9 and Islamic banks (IFRS 9*ISLAMIC) is also negative but statistically insignificant, which suggests that the impact of IFRS 9 does not vary whether a bank is a conventional or an Islamic bank. However, we find that the coefficient of IFRS 9 is significantly negative, which is consistent with our baseline regression results. Regarding control variables, we do not find any impact of firm size on the relation between IFRS 9 and cash holding, which implies that the inverse impact of IFRS 9 does not vary because of size differences. However, we find that cash holding is significantly higher in banks with greater debt ratios. These results suggest that firms hold greater cash to pay their external debt. Cash holding is higher for firms with higher customer deposits, which indicates the bank’s operational efficiency and liquidity strength of the banks concerned. However, we do not find any impact of corporate governance characteristics on the relation between IFRS 9 and corporate cash holdings, which warrants further research.
Regression results of IFRS 9 and cash holding (Conventional banks and Islamic banks)
| Variables | CASH | t-statistic | CASH | t-statistic | CASH | t-statistic |
|---|---|---|---|---|---|---|
| IFRS9 | −0.036*** | [−4.07] | −0.042*** | [−3.80] | −0.045*** | [−4.12] |
| ISLAMIBANK | −0.014 | [−1.39] | ||||
| ISLAMI*IFRS9 | −0.002 | [−0.20] | ||||
| FIRMSIZE | 0.005 | [0.43] | 0.004 | [0.32] | ||
| DEBT | 0.171*** | [4.27] | 0.222*** | [4.67] | ||
| CAPEX | 0.544* | [1.87] | 0.517* | [1.76] | ||
| LOANS2TA | 0.006 | [0.93] | 0 | [0.03] | ||
| DEPOSIT2TA | 0.060*** | [4.33] | 0.070*** | [4.33] | ||
| TIER1 | −0.141 | [−0.82] | −0.081 | [−0.48] | ||
| TIER2 | 0.186 | [0.49] | 0.022 | [0.06] | ||
| ROA | −0.396 | [−0.56] | −0.859 | [−1.11] | ||
| GOODWILL | −6.661*** | [−2.74] | −5.792** | [−2.21] | ||
| SIZE_BOARD | −0.001 | [−0.91] | 0 | [−0.35] | ||
| INDBOD | 0.002 | [0.73] | 0.003 | [0.90] | ||
| SIZE_AC | 0.002 | [1.00] | 0.003 | [1.41] | ||
| INDIRAC | 0 | [−0.04] | 0 | [−0.11] | ||
| DIVIDEND | 0.005 | [0.67] | 0.004 | [0.61] | ||
| OCF | 0.086 | [0.98] | 0.08 | [0.95] | ||
| LNAGE | 0.006 | [0.81] | 0.006 | [0.79] | ||
| ZSCORE | −0.044 | [−1.35] | −0.074** | [−2.12] | ||
| PPE | 0.041 | [0.11] | −0.216 | [−0.51] | ||
| CONSTANT | 0.105*** | [17.26] | −0.075 | [−0.34] | −0.054 | [−0.24] |
| Observations | 232 | 232 | 232 | |||
| R-squared | 0.11 | 0.34 | 0.35 | |||
| ADJ. R-squared | 0.08 | 0.26 | 0.27 | |||
Note(s): *, **, and *** indicates coefficients are statistically significant at 10 percent, 5 percent and 1 percent respectively
Source(s): Authors’ own creation
4.3 Sensitivity analyses
4.3.1 An alternative measure of cash holdings
We use an alternative measure of cash holding based on net cash holding following prior research (Francis et al., 2014). We scale cash holdings by the book value of total assets net of liquid assets. The result of the analysis is presented in Table 5. We find that the coefficient of IFRS 9 is negative and statistically significant at 1%, which is consistent with our main analysis, and it implies that firms decrease cash holdings in post-IFRS adoption periods. Similarly, we run Eq. (2) to see the variance (if any) because of Islamic banks and we do not find any moderating effect, which is in line with our baseline regression results. In sum, we can infer that our results do not suffer from limitations due to measurement differences. Control variables show signs and significance consistent with prior related literature and identical to our main analysis in Table 4.
Regression results (using an alternate measure of cash holdings)
| Variables | NETCASH | t-statistic | NETCASH | t-statistic |
|---|---|---|---|---|
| IFRS9 | −0.052*** | [−3.73] | −0.057*** | [−4.08] |
| ISLAMIC | −0.019 | [−1.50] | ||
| ISLAMIC × IFRS9 | −0.001 | [−0.07] | ||
| FIRMSIZE | 0.006 | [0.39] | 0.005 | [0.29] |
| DEBT | 0.211*** | [4.01] | 0.279*** | [4.48] |
| CAPEX | 0.639* | [1.69] | 0.612 | [1.61] |
| LOANS2TA | 0.007 | [0.82] | −0.001 | [−0.11] |
| DEPOSIT2TA | 0.062*** | [3.63] | 0.076*** | [3.74] |
| TIER1 | −0.17 | [−0.78] | −0.085 | [−0.40] |
| TIER2 | 0.224 | [0.46] | −0.004 | [−0.01] |
| ROA | −0.498 | [−0.55] | −1.119 | [−1.13] |
| GOODWILL | −7.997** | [−2.58] | −6.926** | [−2.06] |
| SIZE_BOARD | −0.001 | [−0.91] | 0 | [−0.34] |
| INDBOD | 0.003 | [0.71] | 0.004 | [0.89] |
| SIZE_AC | 0.003 | [0.98] | 0.004 | [1.43] |
| INDIRAC | 0 | [−0.02] | −0.001 | [−0.10] |
| DIVIDEND | 0.006 | [0.64] | 0.005 | [0.57] |
| OCF | 0.128 | [1.08] | 0.118 | [1.04] |
| LNAGE | 0.008 | [0.83] | 0.008 | [0.80] |
| ZSCORE | −0.06 | [−1.40] | −0.100** | [−2.18] |
| PPE | 0.024 | [0.05] | −0.318 | [−0.57] |
| CONSTANT | −0.09 | [0.32] | −0.061 | [−0.21] |
| Observations | 232 | 232 | ||
| R-squared | 0.32 | 0.34 | ||
| Adj. R-squared | 0.24 | 0.25 |
Note(s): *, **, and *** indicates coefficients are statistically significant at 10 percent, 5 percent and 1 percent respectively
Source(s): Authors’ own creation
5. Further analysis
5.1 Effects of bank size
To reveal the impact of bank size on the effect of IFRS 9 on corporate cash holdings, we divide our sample into larger banks and smaller banks based on the total assets ratio. We use scale variables to minimize the firm size dissimilarity and incompatibility. Hence, we conjecture that larger banks will require more cash holdings to continue their business, including loans to corporate clients and to support the needs of their depositors. Moreover, larger banks invest more in infrastructural development, which deploy more liquid assets compared to smaller banks. However, it is unclear the impact of IFRS 9 on corporate cash holdings of larger banks to smaller banks. To test our hypothesis, we employ Eq. (1) on both samples. Results are presented in Table 6. The first two columns (col 1 and col 2) show the impact of firm size on the effect of IFRS 9 on cash holdings. We find that the coefficient is negative and statistically significant for larger banks, which implies that larger banks hold lower cash upon adoption of IFRS 9. However, we do not find any impact of IFRS 9 on smaller banks, which suggests that smaller banks are less likely to be affected by IFRS 9. Control variables are showing consistent signs and significance.
Regression results of IFRS 9 and corporate cash holdings (sub-sample analysis)
| Firm size (based on total assets) | Firm size (based on total loans disbursed) | Firm age | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Variables | CASH_TA | t-statistic | CASH_TA | t-statistic | CASH_TA | t-statistic | CASH_TA | t-statistic | CASH_TA | t-statistic | CASH_TA | t-statistic |
| IFRS9 | −0.023** | [−2.20] | 0.006 | [0.21] | −0.052*** | [−4.35] | −0.037** | [−2.17] | −0.036** | [−2.21] | −0.018 | [−1.15] |
| FIRMSIZE | 0.004 | [0.27] | 0.018 | [1.06] | −0.023 | [−1.28] | 0.017 | [0.82] | ||||
| DEBT | −0.235* | [−1.96] | 0.065 | [0.80] | 0.082 | [0.72] | −0.075 | [−0.95] | 0.196*** | [3.81] | 0.248*** | [4.54] |
| CAPEX | 0.998*** | [3.17] | 2.939 | [1.52] | 2.769* | [1.67] | 0.425 | [1.15] | 0.427 | [1.21] | −0.184 | [−0.09] |
| LOANS2TA | 0.005 | [0.62] | 0.039 | [1.51] | 0.078*** | [3.66] | 0.024 | [1.39] | ||||
| DEPOSIT2TA | −0.102 | [−1.06] | −0.034 | [−0.67] | −0.014 | [−0.15] | 0.108*** | [4.99] | 0 | [−0.03] | 0.025** | [2.11] |
| TIER1 | 0.213 | [0.92] | −0.626 | [−1.56] | 0.053 | [0.27] | −0.578* | [−1.97] | 0.058 | [0.41] | −1.012** | [−2.65] |
| TIER2 | −0.904* | [−1.80] | 1.788** | [2.15] | −0.574 | [−1.53] | 1.444** | [2.21] | 0.4 | [1.22] | 1.073 | [1.29] |
| ROA | −2.928*** | [−3.02] | 0.641 | [0.35] | −0.364 | [−0.47] | −0.44 | [−0.26] | −1.477 | [−1.33] | 0.568 | [0.57] |
| GOODWILL | −19.181** | [−2.40] | −6.628 | [−1.38] | −13.956*** | [−3.84] | 0 | [0.00] | −7.122*** | [−2.68] | 0 | [0.00] |
| SIZE_BOARD | 0 | [−0.40] | −0.003 | [−1.31] | −0.003*** | [−2.67] | 0.002 | [1.28] | −0.001 | [−0.88] | −0.002 | [−1.38] |
| INDBOD | 0.007** | [2.14] | 0.007 | [0.81] | 0.002 | [0.47] | 0.003 | [0.90] | 0.003 | [0.80] | −0.001 | [−0.07] |
| SIZE_AC | 0.003 | [0.89] | 0.006 | [1.42] | 0.002 | [0.86] | −0.003 | [−0.51] | 0 | [−0.16] | 0.006 | [1.00] |
| INDIRAC | 0.003 | [0.45] | −0.018** | [−2.02] | −0.001 | [−0.09] | −0.006 | [−0.92] | 0.001 | [0.18] | −0.002 | [−0.34] |
| DIVIDEND | −0.002 | [−0.22] | 0.041* | [1.96] | 0.019*** | [2.65] | −0.017** | [−2.17] | 0.005 | [0.57] | 0.012 | [0.92] |
| OCF | −0.063 | [−0.71] | 0.279* | [2.01] | 0.108 | [1.15] | −0.062 | [−0.48] | 0.002 | [0.02] | 0.058 | [0.31] |
| LNAGE | −0.001 | [−0.19] | 0.006 | [0.31] | 0.008 | [0.68] | 0.013 | [1.10] | ||||
| ZSCORE | 0.042 | [0.88] | 0.028 | [0.35] | −0.104** | [−2.30] | 0.183*** | [3.31] | −0.015 | [−0.32] | −0.096** | [−2.47] |
| PPE | −0.219 | [−0.49] | 0.526 | [0.47] | 0.529 | [1.03] | −1.771* | [−1.93] | −0.909* | [−1.79] | 1.619*** | [3.90] |
| Year effects | Controlled | Controlled | Controlled | |||||||||
| CONSTANT | 0.187** | [2.06] | 0.003 | [0.04] | 0.03 | [0.11] | −0.36 | [−1.12] | 0.509 | [1.48] | −0.309 | [−0.76] |
| Observations | 163 | 69 | 148 | 84 | 143 | 89 | ||||||
| R-squared | 0.25 | 0.7 | 0.39 | 0.67 | 0.4 | 0.55 | ||||||
| Adj. R-squared | 0.13 | 0.53 | 0.27 | 0.55 | 0.28 | 0.4 | ||||||
Note(s): *, **, and *** indicates coefficients are statistically significant at 10 percent, 5 percent and 1 percent respectively
Source(s): Authors’ own creation
5.2 Effects of banks' loans and investments
In this section, we attempt to capture whether the impact of IFRS 9 on cash holdings varies between banks with higher loans and investments over banks with lower loans and investments. This is because issuing loans is one of the pivotal functions of banks and it involves risks as well. Based on the deposit to total assets, we divide our entire sample into two groups (larger banks and smaller banks), and we attempt to explore the impact of IFRS 9 on these two groups of banks using equation (1). Results are presented below in Table 6. We find that the negative relation between IFRS 9 and cash holding is highly pronounced in larger banks relative to smaller banks. Our results suggest that larger banks hold less cash upon the adoption of IFRS 9. Other control variables are showing consistent signs and statistical significance.
5.3 Effect of bank age (older vs. young banks)
Prior research shows that older banks are more sustainable than young banks (DeYoung et al., 1999), and compared to young banks, older banks get funds easily and with easier terms and conditions from savers. Taking this tension, we attempt to explore the impact of IFRS 9 on older banks over young banks. We classify all banks into these two groups based on the median age (the difference between the current year and the listing year with the stock exchange). We employ Eq. (1) for these two samples. Our results show that the negative relation between IFRS 9 and corporate cash holdings is highly pronounced in older banks compared to young banks. Our results suggest that older banks keep less cash in post-IFRS 9 adoption periods. This result can also be attributed to the notion of investment efficiency or reinvestment in profitable channels. On the other hand, smaller banks require more cash to survive in the market, and they are less likely to be affected by IFRS 9.
All other control variables show a consistent sign and significance, which is in line with our baseline regression regarding IFRS 9 and corporate cash holdings.
6. Conclusion
This paper is intended to explore the impact of IFRS 9 on corporate cash holdings. While limited prior research on IFRS is conducted to test its impact on non-financial sectors, our paper focuses on the banking sector as IFRS 9 heavily affects the financial sector. We tested our conjecture by taking a sample of publicly listed banks in Bangladesh where IFRS 9 was mandatorily enforced in 2018. Corporate cash holding is one of the salient issues in Bangladesh, as the majority of banks are privately owned and managed. Bangladesh provides an excellent setting to test the impact of IFRS 9, as all of the listed banks adopted IFRS 9 in the same year. Nevertheless, all publicly listed banks are subject to the same regulators, and they are following the same set of corporate governance guidelines. However, we notice that many banks are suffering from liquidity and customers are unable to withdraw their own deposit money on demand [2]. Moreover, IFRS 9 suggests a set of provisions that all banks must follow. Particularly, IFRS 9 affects firm-level provisions regarding loans and advances, revaluation reserves, investments and other comprehensive income. However, none of the prior research attempts to explore the impact of IFRS 9 on banks' cash holdings in an emerging market and we attempted to fill this void.
Using a sample of Bangladeshi publicly listed banks, we find that cash holdings are significantly lower in post-IFRS 9 periods compared to pre-IFRS 9 periods. We also investigate whether the above negative relation varies between Islamic banks and conventional banks in the same setting. However, we do not find any moderating role of Islamic banks on the said nexus. Furthermore, we conduct several channel analyses to explore determinants of the above negative relation between IFRS 9 and corporate cash holdings. We find that corporate cash holdings are significantly lower for larger banks and older banks in post-IFRS 9 periods compared to smaller and younger banks. However, we document the negative impact of IFRS 9, but no significant difference due to the bank’s size in terms of loans and advances.
The present study contributes to the extant literature of IFRS on corporate cash holdings as the evidence of IFRS 9 in the banking sector is scant, particularly from the emerging country perspective. Second, this study brings a scenario of the impact of IFRS 9 on both Islamic banks and conventional banks, which gives an added novelty to the existing IFRS research. Third, our paper shows how the management of a bank company focuses its attention on fund management and cash management when an international accounting standard comes into existence. Fourth, researchers will benefit from the present study as we show how a single accounting standard affects the firm liquidity position of banks in a developing country. Unlike developed countries, emerging countries suffer from institutional complexity, limited relevant expertise in management, family control and related party transaction crises. Taking all of these characteristics, our study will be highly relevant for the regulator, academics, researchers and other stakeholders who are directly or indirectly associated with the banking industry.
Lastly, we mention some caveats of our paper. First, the present study does not show any causal relationship between IFRS 9 and corporate cash holdings. Instead, we depend on the association to explore the possible relation between IFRS 9 and corporate cash holdings. Second, we limit our sample periods from 2015 to 2022, but such a study should cover at least two regimes as the country experienced more than a decade with the same government. We believe future studies can take larger samples covering more than one political regime to explore if there is any impact of government change on the said relation. Third and finally, we are unable to rule out the possibility of the impact of the changes in the corporate governance code 2018. Future researchers can incorporate the moderating effect of corporate governance code on the impact of IFRS 9 on corporate cash holdings in the same setting.
Note
The supplementary material for this article can be found online.
