“A longitudinal study of corporate social responsibility expenditure and ownership structure of financial firms”

There is a dearth of longitudinal studies of corporate social responsibility expenditure (CSRE) and corporate governance in Bangladesh, which has been the impetus for this study. The study aims to identify the relationship between ownership structure and CSR expenditure. The empirical study considered a longitudinal period of 2007–2019 of listed financial firms (banks and non-banking financial institutions) of Bangladesh. The final sample consisted of 461 firm-year observations for 53 firms. The study incorporated a set of theories, including agency cost theory and stakeholder theory. The study applied the ordinary least square (OLS) regression technique to test hypotheses. The results of multiple regression analysis showed that foreign ownership and managerial ownership contribute positively and significantly to CSRE. However, the study did not document any relationship between institutional ownership and CSRE. The study used rigorous and alternative measurement techniques to further verify the findings. It was concluded that value creation from CSRE is highly dependent on the ownership structure of financial firms. The empirical study has significant theoretical and managerial implications.


INTRODUCTION
The nature of CSR and environmental investment and expenditure largely depends on the top management priority; it is also more evident when the sample consists of developing countries.Investors are the most influential part of any organization to take any sort of decision.Not only investment, but also routine business management decisions depend on the corporate policy pursued by different shareholders and investors.The monetary decision of CSR mostly depends on the corporate governance elements.Stakeholder engagement and regulatory compliance also influenced CSR expenditure decisions.Firms in developing countries face limited stakeholder pressure, and there is also a lack of regulatory compliance (Bae et al., 2018(Bae et al., , 2021;;Rashid, 2021;Bose et al., 2020a;Masud et al., 2019Masud et al., , 2018aMasud et al., , 2018bMasud et al., , 2017)).In the presence of weak corporate governance, the financial system of emerging countries like Bangladesh has a conflict between shareholders and agents due to the benefits and short-term investment decisions.It is essential to examine whether the stakeholders and agency theory are the best approach to explain the relationship between ownership structures and CSR expenditure.
Most of CSR studies highly focused on disclosure performance (non-financial) rather than monetary-based performance (financial).The prior study used content analysis and third-party rating to evaluate the CSR performance of firms.Nevertheless, the non-monetary nature of content-based disclosure and rating are not enough to measure the monetary-based CSR performance of firms (Bose et al., 2020b).The most recent study by Khan et al. (2021) raised the issue of 'green washing' or 'authenticity effort' of the sustainability reporting quality of financial firms.CSR expenditure is considered a value-driven approach that allures management to invest more in CSR activities to inform various stakeholders (Bose et al., 2020b).Khan et al. (2021), Bose et al. (2020b), and Bhattacharyya and Rahman (2020) empirically proved that CSR expenditure enhances firm growth, stability, and quality of information because of high-value creation and keeping the commitment.As already mentioned, most of CSR research focused on non-financial performance measurement rather than financial attributes.Prior studies of Khan et al. (2021), Bose et al. (2020b), Bhattacharyya and Rahman (2020), Feng et al. (2015), Manchiraju and Rajgopal, (2017), Dharmapala and Khanna (2018), Kapoor and Dhamija, (2017), and Brammer and Millington (2008) have examined the actual CSR expenditure performance of firms.Furthermore, the above mentioned actual CSR expenditure studies examined market value (Bose et al., 2020b), stock return (Bhattacharyya & Rahman, 2020), quality information of sustainability reporting (Khan et al., 2021), shareholder value (Manchiraju & Rajgopal, 2017), CSR compliance (Dharmapala & Khanna, 2018; Kapoor & Dhamija, 2017).No study considered the impact of actual CSR expenditure on corporate governance.This study was undertaken as an attempt to fill this literature gap.
The study is designed to ask a question under what circumstances corporate governance element (ownership structure) decides to use CSR expenditure.To explore the research question based on the literature gap, the study is designed to investigate the extent of the ownership structure impact on CSR expenditure.To the authors' knowledge, this is the first study that considers CSR expenditure to evaluate the relationship between ownership structure and CSR expenditure.Bangladesh has recently entered into the list of developing countries.Its financial sector is now booming and undergoing expansion, with several local and international banking institutions involved in the field.As new banks are permitted to enter Bangladesh's financial market, competition in the banking sector has intensified over time.It is also expected that future competition between public and private banks will be fierce (Rahman et al., 2021).Moreover, the grounds for considering Bangladeshi financial firms are as follows: 1) the financial sector is the most regulated in the country compared to the non-financial sector (Bae et al., 2021); 2) the central bank of the country promulgated CSR regulation since 2008 for the financial sector that ensures availability of CSR expenditure data of financial sector; 3) Bangladesh is the most emerging economy in the world, while CSR expenditure is very helpful to mitigate poverty alleviation (World Bank, 2017), along with the achievement of remarkable progress on the Sustainable Development Goals (SDGs).
To investigate the relationship between ownership structure and CSR expenditure, the study incorporated a set of hypotheses, including listed financial sector firms of Bangladesh.Finally, the study documented 461 firm-year observations during 2007-2019.H1: CSR expenditure is positively and significantly associated with foreign ownership.

Institutional ownership and CSR expenditure
Institutional owners are considered powerful controllers in the corporate governance elements because they hold a large part of the share (Rashid et al., 2020) H2: CSR expenditure is positively and significantly associated with institutional ownership.

Director ownership and CSR expenditure
Directors of a firm are mandated for key decisions in the business; therefore, they have active and influential access to the investment decision.Moreover, directors have agency conflict with the owners, so that they like to show business performance in the short term (Khan et   In addition, certain alternative variables were employed to test the study's robustness.The measurement of all variables is given in Table 2.

Data estimation model
The study used unbalanced panel data.As the data of this study are static in nature, general regres- where CSRE = a ratio of corporate social responsibility expenditure scaled by total revenue by each sample firm; α 0 = the constant; ε it = error terms; β 1 to β 11 = the coefficients of the variables defined in Table 1, and 'i' and 't' = the number of banks and periods, respectively.

Descriptive statistics
The mean and standard (SD) are used to measure the overall performance of Bangladeshi listed financial institutions.Table 3 presents the descriptive data of the sample organizations.The CSRE has a mean of 0.005, which indicates sample firms spend on average one to two percent of total revenues as CSRE.CSR expenditure of financial firms is different from each other and CSRE is very limited in amount.Among the ownership structure, most of the shares are controlled by the directors as the percentage of their shareholding is 38.57percent, while institutional and foreign ownership is 16.81 percent and 3.31 percent, respectively.There is a substantial possibility that directors can influence the board in all decision-making.The average board size is 13 percent, where the average size of independent and foreign directors is 0.17 and 0.23, respectively.

Correlations matrix
The correlation coefficient between any pair of variables utilized in this investigation is shown in Table 4.It shows whether the correlation between the variables is substantial in a negative or positive way.Table 4 shows that CSRE is positively and significantly associated with directors' ownership, the board size, independent directors, firm size, leverage, and advertising expense.On the contra-CSRE is negatively significant to ROA; it indicates higher financial performance, lower contribution to CSRE.These results may help in further regression analysis.

Bivariate analysis
Multicollinearity is a critical issue to verify the correctness of a model, since one predictor variable in Table 5 shows that all values below the threshold of 10 indicate that there is no multicollinearity among the variables in the study (Hair et al., 1995).

Regression results
The association between ownership structure and CSRE performance is seen in Table 5. Static els such as pooled OLS, fixed effect, and random effect models were used in the study.The explanatory variables explained nearly 44% of the variation in the dependent variable "CSR expenditure," according to the adjusted coefficient of determination (Adj.R 2 ).The study can take the unmistakable conclusion that this empirical model fits the data well as the p-value of the F-test is statistically significant (17.37, p < 0.01).The pooled OLS model shows a statistically significant impact of ownership structure on CSRE performance.However, endogeneity in ownership pattern is not a concern for Asian firms; Bangladeshi firms are relatively stable over time.Additionally, the present study has used the dummy to deal with the heterogeneity issue by using year-fixed effects.than 50 percent of shareholding investors being unable to influence firm performance positively (Al-Gamrh et al., 2020).
In the ownership structure, the last variable is DRTS, which is positively and significantly associated with CSRP (p < 0.01).Hence, the study accepts H3, and the result is supported by the studies of Jahid et al. ).The study documented that the average DRTO share is 39 percent that can control the investment decision over CSR expenditure.Moreover, prior studies also argued that higher shareholding by the director reduces CSR performance because of time horizon, personal interests, and lack of stakeholder engagement (Dam & Scholtens 2013).Hence, the study tends to believe that more DRTO helps management consider CSRE as a strategic issue rather than philanthropic that will carry more opportunities and eliminate possible threats.CSR expenditure is a very new concept in Bangladesh, which enjoys

Robustness tests
To prove the authenticity, the study incorporated additional sensitivity analysis.If the main OLS (Table 5) result is similar to the additional analysis (Table 6), the findings are robust.For robustness check, the study used alternative variables both for dependent and independent variables.First, the study used an alternative dependent variable measured as a ratio of total CSR expenditure scaled by total assets (column 2) and found simi- The study tested ownership structure variables, and the result pointed out that the nature of CSR expenditure of the country's financial sector firms majorly depends on the corporate governance elements.The study reported that foreign shareholding significantly influenced more CSR expenditure.
The study believes that cultural differences, best industry practice experience, and more ingrained attitudes influence foreign investors to take positive action of higher CSR expenditure.Moreover, a foreign investor likes to share the host country's social and environmental experiences in the new business environment because of engaging stakeholders and creating a positive impression of the business.Bangladesh is an emerging economy, therefore, policymakers and business management should invite more foreign investors to change the prevailing business environment and enjoy competitiveness.On the other hand, the study did not find any significant value of institutional shareholding that is very surprising but consistent with the prior study.Institutional investors invest in the long-run project, but in Bangladesh, long-run business products are absent (Rashid, 2021).It is also noteworthy to mention that the capital and money market of the country is run by short-term investment funds, including mutual funds and superannuation funds (Rashid, 2021).Moreover, the shareholding pattern of institutional investors of the country is minimum, which would be a possible reason for ignoring CSR expenditure.The study also used managerial ownership and found a positive association with CSRE.Directors hold most of the share that makes them controllers of the management and business.A higher shareholding pattern helps them to invest higher CSR expenditure for gaining business reputation, more visibility, and market growth.CSR expenditure helps to create value to the business, therefore director ownership likes to establish strategically CSR in the decision-making for future benefits.Director-owners control the business, therefore, they want to keep the commitment, actual CSR expenditure helps them to revise CSR policy and regulations.Moreover, to check the manager's CSR expenditure function, it is also required to establish a strong regulatory framework because higher shareholding inspires the director to involve political CSR for showing personal power and implementing political party's agenda (Uddin et al., 2016).

CONCLUSION
The study has significant theoretical and managerial contributions.To the best of the authors' knowledge, this is the first study of the impact of ownership structure on CSR expenditure.Therefore, the study has a significant contribution to corporate governance theory.CSR expenditure can be utilized to measure the performance of the corporate governance element.The monetary nature of CSR function is the new mechanism to evaluate corporate management commitment and performance.Theoretically, the study contributed to the agency cost and stakeholders.Higher CSR expenditure can be engaged to reduce agency conflict because management (agent) will be strictly evaluated by the principal (own-er).It will be also very fruitful to the owner to check each monetary CSR function performed by the management (i.e.audited by a third party, assurance statement).On the contrary, stakeholder theory contributes to the stakeholder engagement mechanism through CSR expenditure.Different stakeholder groups can scrutinize the actual amount of money used by the name of CSR.Visible CSR expenditure efficiently and effectively communicates between stakeholders and management that will ultimately enhance market reputation, transparency, growth, stability, and performance.Hence, CSRE can be used as an effective mechanism between stakeholders and management for enhancing trust and confidence.Moreover, the study has significant managerial and policy implications.Higher CSRE increases profitability and firm value, therefore, strong regulation is inevitable in this sector.Financial firms are utilizing CSRE for value creation that could be a role model for the non-financial sector.Business management should use the strategic CSR concept rather than philanthropic.Policymakers should promulgate a holistic mandatory strategic CSR framework to all sectorial firms.
Despite significant contributions, the study is not without limitations.Firstly, the study only considered listed financial firms.Secondly, it omitted some influential governance variables and the impact of political CSR expenditure.Thirdly, the study did not incorporate appropriate theoretical setting to best explain corporate governance and CSR expenditure (i.e.organization identity theory).Lastly, the study ignored market-driven variables and organizational regulatory forces.Future studies should incorporate financial and non-financial firms to investigate all corporate governance elements.Moreover, a future study can be taken into consideration to investigate regulatory framework, market forces, and profitability impact on CSR expenditure.
(2020), Jia and Zhang (2012), and C.-Y. Huang and S. W.-L. Huang (2009).Higher managerial ownership reduces agency cost and encourages management to invest more in CSR expenditure because of managerial and corporate reputation (Bingham et al., 2011), closely tied to the firm as a general shareholder (Goergen & Renneboog 2010), encourages strict monitoring and controlling to any socially irresponsible investment and behavior (Goergen & Renneboog, 2010), and opportunities to establish business goals (Li & Zhang, 2010

Table 1 .
). Sample selection (1)p://dx.doi.org/10.21511/bbs.17(1).2022.032.2.Variable measurement2.2.1.Dependent variableThe study used CSR expenditure (CSRE) as a dependent variable to find the impact of ownership structures on it.Bangladesh Bank (BB) guided eight areas, including education, idly explains the motive of the management to use money in this area.Following the study ofKhan  etal.(2021),Baeetal. (2021), Bose et al. (2020b), and Bhattacharyya and Rahman (2020), the study incorporated CSRE to explore the research goals.To measure the CSRE ratio, the study scaled total CSR expenditure by total revenue.CSRE data is managed from BB's quarterly published2.2.2.Independent variablesThe study used different corporate governance elements as independent variables, including ownership structure.The study used foreign ownership (FRWO), institutional ownership (INSO), and director ownership (DRTO) following the studies by Jahid et al. (2020), Masud et al. (2018a, 2019), and Bae et al. (2018).The definition and measurement of corporate governance data are presented in

Table 2 .
Variable definitions and measurements

Table 3 .
Descriptive statistics Variable definitions and measurements are given in Table2.

Table 4
. As a result, there is no concern with multicollinearity in this study.In addition, the study used the variance inflation factor (VIF) test to see if the model has a collinearity problem.

Table 6 .
Sensitivity analysis results to the baseline model.Second, the study took FRND as an alternative proxy of FRNO (column 3) and reported a positive and significant association (p < 0.01), which is also similar to the baseline model.Third, the study considered ROE as an alternative proxy of ROA (column 3) and documented similar findings of the baseline model.Finally, the study replaced firm REVENUE instead of FSIZE, and the results remain unchanged.Additionally, every model's results are in line with the baseline model that proves the study as robust.
Note: FRND presents the number of foreigners on the board; REVENUE indicates the log of total revenues of the individual firm; ROE represents net income divided by total equity.Standard errors in parentheses; * p < 0.10, ** p < 0.05, and *** p < 0.01.lar