“The association between Corporate Social Responsibility Disclosure and accounting-based financial performance: a Kuwaiti evidence”

The current study aims at extending prior accounting research on the association be-tween Corporate Social Responsibility Disclosure (CSRD) and Corporate Financial Performance (CFP) using a sample of listed firms on Kuwait Stock Exchange (KSE) from 2011 to 2012. It conducts a regression analysis to investigate the association be-tween CSRD and CFP, as well as investigates the impact of firm size, leverage, and industry affiliation as the key determinants suggested by prior research on the level of CSRD. The results of the present study reveal that both CFP and firm size have significant positive associations with CSRD, whereas, in contrast, firm’s leverage and firm’s industry affiliation show non-significant associations with CSRD.


INTRODUCTION
For the last 20 years, Corporate Social Responsibility Disclosure (CSRD) has attracted a growing admiration and respect among academics, as well as professionals (e.g.Chang et al., 2017).Elias and Epstein (1975, p. 36) pointed out that "CSRD involves the reporting of aspects related to the way a business entity is involved in social activities Similarly, Gray et al. (1996) defined it as a communication process that firms use to communicate the social impacts of their economic actions to specific groups within their society.It shows how business entities as members of the society address both social and environmental issues of concern to the surrounding communities in their operations and activities (Kok et al., 2001).The association between CSRD and Corporate Financial Performance (CFP) was an issue that has drawn the attention of several accounting researchers (e.g., Muttakin et al., 2015;Chang et al., 2017;Bani-Khalid et al., 2017).Investigating the CSRD-CFP association has been an appealing research issue, as obtaining an understanding about this association is expected to be instrumental for the corporate management, as well as the different groups of the firm's stakeholders (Simpson & Kohers, 2002).In a review of prior CSRD-CFP, Gray (2006) suggested that empirical findings reported about this association are mixed, leaving only indecisive conclusions about this research issue.In addition, prior related research suggests that investigating the CSRD-CFP association is knotty both theoretically and methodologically (e.g.Carroll, 2000).Hence, further investigation of this association is apparently needed.
The objective of the current study is to answer two main research questions.First, whether there is a significant association between CSRD and CFP for a sample of Kuwaiti listed companies.Second, whether CSRD in Kuwait is affected by the same key drivers documented in prior CSRD research.As indicated earlier, the research questions of interest in the current paper may have been examined considerably in prior research and across different countries.Empirical results reported about these questions, however, are still not conclusive (Bani-Khalid et al., 2017).In addition, much of previous empirical research examining factors influencing CSRD has been conducted in the context of Western or East-Asian countries' markets, with very few studies carried out to examine such issues in the context of other parts of the world.Furthermore, no research, the authors are aware of, has empirically investigated the CSRD-CFP relation in the context of the emerging Kuwaiti market.The current study, therefore, intends to expand this line of accounting research by providing empirical evidence about this relation from a setting that has not been explored previously, where corporate governance regulations has been almost absent, and where corporate ownership structure in listed firms is featured by high concentration (Al-Fraih et al., 2012).In particular, this study attempts to contribute in filling the extant shortage in the international literature by providing further empirical evidence about (i) whether there is an association between CSRD and CFP in the context of the Kuwaiti market, and (ii) whether CSRD in Kuwait is incfluenced by some key CSRD determinants, namely firm size, leverage, and industry affiliation.The current study, therefore, contributes to the international literature by reporting empirical evidence about these research questions in the context of an unexplored emerging market.The results reported in the current study reveal that CSRD in Kuwait is significantly associated with CFP as well as firm size.The results reported, nonetheless, financial leverage nor industry affiliation have any significant associations with CSRD.The findings offered by this study are anticipated to be interest to accounting and business researchers, and to be of practical implications for corporate management, investors and different interested parties.The structure of the paper proceeds as follows.Firstly, section 1 presents an overview on the study's theoretical framework.Section 2 reviews the related literature.Secondly, section 3 offers some details on the research method.Section 5 shows the empirical analysis of the current study.Finally, last section reports the study's conclusions.

THEORY AND BACKGROUND
1.1.Theoretical framework of the study CSRD practices have been studied in light of a number of theoretical perspectives such as stakeholder theory, agency theory, legitimacy theory and institutional theory.However, stakeholder and legitimacy theories often have been used to explain why business entities engage in CSRD.

Stakeholder theory
Solomon (2010) posits that "stakeholder theory could be regarded as a conceptual mixture from different disciplines that renders interesting sociological and organizational interpretations for corporate behaviors.Stakeholder theory gives a great deal of attention to the company's stake-holders and argues that a company should give special attention to meet their interests and needs (Maignan & Ferrell, 2004).From the perspective of stakeholder theory, not only corporate shareholders are affected by firm's activities, but also other different groups (Freeman, 1999).The stakeholder perspective argues that besides shareholders, other constituencies are influenced by firm's activities (Freeman, 1999).The basic idea here is that the success of the company and its continuation in its activities depends mainly on how the company manages its relations with these different groups of stakeholders (Freeman & Philips, 2002).

Legitimacy theory
In line with stakeholder theory, legitimacy theory argues that firms' survival depends significantly on their adherence to the boundaries of what the society considers socially acceptable behaviors" (Mousa, 2010).Consequently, a company needs to confirm its legitimacy by maintaining a good re-lationship with its stakeholders to keep their endorsement and support for a company's goals and activities (Elsbach & Sutton, 1992).It is argued that stakeholders who possess greater influence and perseverance will know more about organization's CSR initiatives than other less influential stakeholders (Peloza & Papania, 2008).Legitimacy theory gives a corporate management the chance to handle CSRD to sway and distort the company's reports in a manner, which is favorable to the company (Kokubu et al., 1994).It can be argued that both legitimacy and stakeholder theories are complementary to each other, as they provide an answer to the fundamental question of why firms provide CSRD.While legitimacy theory argues that firms are engaging in CSRD to maintain their legitimacy and to continue their activities in society through obtaining the acceptance of stakeholders, the other theory, stakeholder theory, points out that firms are involving in CSRD to meet the interests and needs of stakeholders.

The Kuwaiti stock market
The Kuwaiti stock market is indeed one of the oldest equity markets in the Arab world.The enormous upsurge in the Kuwaiti economy following the production of oil in the mid of the 20th century has resulted in a huge increase in commerce and business activities in Kuwaiti.This escalation of the volume of business was accompanied by a dramatic upsurge in the number of companies operating in the market, and has resulted in a need for a market where investors can raise and exchange equity.

CSRD and CFP
Literature on the association between CSRD and CFP has offered different approaches with different theoretical foundations.However, mixed and conflicting results are reported.Table 1 shows some of these studies.
A number of studies suggested a positive association between CSRD and CFP with different explanations.For example, proponents of the "social impact hypothesis" of the stakeholders theory (e.g.Alturki, 2014) suggest that when a firm commits itself to meeting the needs of non-owner stakeholders (e.g.customers, employees, etc.), this will lead to enhancing its image and reputation in the society, which will consequently be reflected in its financial performance.This notion is backed by empirical evidence that firm's devotion to (and disclosure of) socially responsible activities and actions leads to lowering its cost of capital, and, in turn, leads to improving its financial results, and vice versa (Dhaliwal et al., 2011).This way, CSRD can be seen as a channel that companies use to communicate their dedication to meeting stakeholders' expectations and demands.Another possible reason suggested for the positive association between CSRD and CFP is that companies that commit themselves to socially responsible behaviors are in general companies with capable management, and hence are expected to perform well financially as well.In fact, the CSRD literature includes research (e.g.additional explanation suggested for the CSRD-CFP positive association is that companies that devote some of their resources for CSRD activities are generally companies with superb CFP (Preston & O'Bannon, 1997).Moreover, some authors (e.g.Milne, 2002) suggested that the positive association between CSRD and CFP is because firms with large profits generally have a greater incentive to disclose their CSRD activities in an effort to divert general public attention and to avoid any potential political intervention away from their profits.
On the other hand, researchers who suggested an inverse association between the level of CSRD and CFP basically argue that a high level of CSRD means more amounts of firm's economic resources spent on CSRD activities, and hence lower amounts of profit (Salzmann et al., 2005).Moreover, research suggesting a negative relation between CSRD and CFP includes also studies (e.g.Preston & O'Bannon, 1997) that proposed the "managerial opportunistic behavior" hypothesis and suggested that corporate management interested mainly in serving its own interest is expected to reduce expenditures on (and disclosure of) CSRD activities when the firm is doing well financially in order to maximize firm's reported profits and hence executive compensation.This hypothesis continues to suggest that when the firm is financially experiencing a bad time, firm's management tend to "opportunistically" increase CSRD in order to divert public attention from its bad performance.
Based on previous discussion, the association between CSRD and CFP has reported mixed results.
Several studies report a significant positive association between the level of CSRD and CFP (e.g., Simpson & Kohers, 2002;Khan et al., 2013).Other studies (Aras et al., 2010), however, could not report any significant association between CSRD and CFP.Although empirical results reported in previous research about the CSRD-CFP association are mixed, these research findings are mostly suggestive of a positive association between the CSRD level and CFP.Therefore, the study's first hypothesis is specified as follows: H1: There is a significant positive relation between the amount of CSRD and CFP.

Firm size
The CSRD literature includes a number of theoretical bases offered to explain the effect of firm size on its CSRD.Since larger firms are more visible to social activists groups, they are expected to H2: There is a significant positive relation between firm size and the amount of CSRD.

Leverage
Several previous CSRD studies ( Wallace et al., 1994;Khlif & Souissi, 2010) showed that firm's leverage is significant in explaining the amount of CSRD by firms.Empirically, however, the directional impact of this variable on the CSRD is still unclear.On the one hand, some studies (e.g., Haniffa & Cooke, 2005) suggested that high leveraged firms have a greater incentive to provide more CSRD, as they need to legitimize their actions.Other studies (e.g., Purushothaman et al., 2000), however, suggested that firms with high leverage are expected to have very close ties with their creditors, and, therefore, are expected to have other channels for disclosing to them their CSRD activities and actions.In view of the aforementioned opposite arguments and mixed empirical results, no prediction is made about the direction of the relation between the amount of CSRD and firm's leverage.The current study's third hypothesis, therefore, is as follows: H3: There is a significant association between firm's leverage and the amount of CSRD.

Industry affiliation
The voluntary disclosure line of research in the accounting literature includes studies (e.g., Dye and Sridhar, 1995) showing that the amount of firms' voluntary disclosure is considerably affected by the characteristics of the industry they belong to.Prior CSRD research (e.g., Cormier et al., 2005) suggested that firms' decision with regard to the amount of CSRD is significantly influenced by their industries' characteristics.Moreover, empirical results documented in previous CSRD research (e.g., Kotonen, 2009) reveal a significant association between the amount of firm's CSRD and its industry affiliation.For example, some studies (e.g., Reverte et al., 2009) found that firms that belong to "environment sensitivity" industries tend to make considerably high amounts of CSRD than other firms.Industries with high environment sensitivity are industries with manufacturing operations that potentially have harmful impacts on the environment.In general, these industries include, for example, oil, mining and chemical industries.Prior related research (Clarke & Gibson-Sweet, 1999) showed that firms belonging to these industries tend to provide more CSRD about environment, health, and safety.Empirical evidence was documented in previous research, therefore, the amount of CSRD is expected to depend greatly on the environment sensitivity of their industry.The current study's forth hypothesis, therefore, is specified as follows: H4: There is a significant positive association between the company's industry affiliation and the amount of CSRD.

CSRD measurement (dependent variable)
As specified earlier, CSRD, the dependent variable in the model, is a measure of the amount of CSRD made by firms in their annual reports.To measure this CSRD variable, an index was calculated based on each firm's score on a checklist of 21 items pertinent to CSRD information disclosed in the annual report (see Appendix for CSRD items).This checklist of CSRD items is a modified version of a checklist used by Hackston and Milne (1996) and several other CSR studies (e.g., Aras et al., 2010).To measure the value of the CSRD index, the current study followed a procedure similar to that used in prior research (i.e., Muttakin & Khan, 2014).Specially, for each observation, any item in the CSRD checklist is given the value of 1 if disclosed by the firm in any part of its annual report, and 0 if not.For each firms, the CSRD index is then obtained by determining the ratio of the total score of the disclosed CSRD items recorded for each firm to the maximum possible score, which is 21.Hence, the CSRD index is calculated as follows: where jindex CSRD -CSRD index for j-th firm, j n - total possible number of CSRD items for j-th firm, ij x -1 if the i-th item of the CSRD checklist is disclosed by firm j, and 0 if not.

Independent variables (explanatory variables)
Prior studies that examine the CSRD As Table 2 shows, the sampled firms come from 12 industries according to KSE industry-sector clas-To test H4, however, the sampled firms were divided into two main groups according to their potential environmental sensitivity.In particular, the ENVIR dummy variable is added to the research model to represent impact of firm's industry affiliation on the amount of CSRD.This variable takes the value of 1 if the firm is an oil and gas, petrochemicals, or a manufacturing firm, and 0 otherwise.

Control variables
Prior CSRD research (e.g., Cormier et al., 2005) suggested that old firms tend to provide more CSRD than newly established firms.Hence, the AGE variable is added to the regression equation to control for the potential impact of firm's age on the amount of CSRD.In addition, prior related research (Walden & Schwartz, 1997) suggested that CSRD (i.e., environment-related information) may change from one year to another.Therefore, the YEAR dummy variable is included in the regression equation to control for the possible impact of year differences.This variable takes the value of 1 if the firm's annual report is for the year 2012, and 0 if the annual report is for the year 2011.

Study sample
The study sample consists of a total of 114 firm-year observations.The data set pertains to 2011 and/or 2012 annual reports of a randomly selected sample of 81 firms listed in the Kuwait Stock Exchange (KSE).Table 2 shows the number of firms included in the sample, and the different KSE sectors the sampled firms are affiliated to.As shown, the sampled firms represent 12 sectors, including banks (11 firms), financial services (23 firms), insurance (3 firms), real estate (17 firms), oil and gas (3 firms), industrial (9 firms), basic materials (2 firms), consumer goods (2 firms), telecommunications (3 firms), healthcare (1 firm), and technology (1 firm).Content analysis was performed on 114 annual reports of firms to extract the data needed for the analysis of interest in current study.In particular, information related to CSRD was hand-collected from the different parts of annual reports, including chairman's statement, directors' report, and supplementary notes.In addition, the financial data needed for the study's analysis were obtained directly from the financial statements' section of the annual reports.The size of the study's sample is fairly reasonable, since the companies included in the study's sample constitute approximately 40% of companies listed on KSE.

Model specification
A regression model is developed based on prior related research (e.g., Simpson & Kohers, 2002;Khan et al., 2013) to investigate the study's research hypotheses.In particular, the following OLS regression is used to examine whether the amount of CSRD is related to the various explanatory variables specified in the research hypotheses: where CSRD -score index of corporate social re- sponsibility disclosure, ROA -the company's net income divided by total assets, SIZE -natural log of the company's total assets, LEV -the com- pany's total liabilities divided by its total equity, ENVIR -dummy variable taking a value of 1 if the company belongs to oil and gas, petrochemicals, or industrial sector, and 0 otherwise, AGE -the number of years since the firm's founding, YEAR -one if the annual report is for 2012 fiscal year, and zero if it is for 2011, ε -error term.
Numerous measures have been used to proxy for CFP.Return on Assets, ROA, however, was one of the most used measures of firm's CFP in previous related research (e.g., Berman et al., 1999;Simpson & Kohers, 2002).According to Jensen and Meckling (1976), corporate agency-related costs increase when the size of the firm increases.Therefore, larger firms are more likely to disclose more information related to their operation and activities than smaller firms.Previous accounting and CSR research (e.g., Murali & Welch, 1989) has used several measures to proxy for firm size.Much of this research, however, have used firm's total assets as a surrogate of firm size.The current study uses total assets as a measure of firm size, and the SIZE variable is, therefore, included in the research model.According to several studies (e.g., Bellaoui & Karpik), the amounts of CSR disclosure is negatively related to firm's financial leverage.The LEV variable is, therefore, added to the equation as a proxy for firm's financial leverage.Previous research (e.g., Sethi, 1976) posits that CSR disclosure is more likely for oil and energy producing firms.As Sethi (1976) suggests, these firms use CSR disclosures as "advocacy advertising".Accordingly, the ENVIR dummy variable is added to the equation as a surrogate for firm's affiliation to the oil and energy industry.

Descriptive statistics
Descriptive statistics are shown in Table 3.As this table shows, firms comprising the study's sample vary from rather small firms, with total assets of KD 5,851,063, to considerably large firms, with total assets of about KD16,424,487,000.Table 3 reveals that the average CSRD score for firms included in the sample is about 0.19.This means that, on average, each firm disclosed about 4 of the 21 CSRD checklist items in its annual report.Table 3 shows that the average ROA for the sampled firms is about 3.8%, while the average leverage ratio is about 2.13.Table 3 also shows that the average age of the sampled firms is around 30 years, and that about 18% of firms included in the sample were from "environmentally sensitive" industries (e.g., oil, petrochemicals, and industrial firms).As shown in Table 3, about 58% of the study's observations pertain to 2012 annual reports, while about 42% of the observations are related to 2011 annual reports.

Univariate analysis
Pearson correlations for the variables included in the regression equation are shown in Table 4.As Table 4 demonstrates, CSRD score is positively and significantly correlated with firm size (p-val- As Table 5 shows, the R-square of the regression model is 0.140, and the F-statistics is 2.892 (p-value = 0.012).The regression results show that the CSRD score is positively and significantly related to firm's ROA (p-value = 0.019), as well as firm's size (p-value = 0.026), which provide support to H1 and H2.In particular, these empirical re-  sults show evidence of a significant and positive association between the amount of CSRD made by firms listed on KSE and their financial performance.Moreover, the results in Table 5 show that the amount of CSRD made by firms listed on KSE is significantly and positively related to firm size.This result means that, in general, larger firms listed on KSE tend to provide more CSRD than smaller firms.The reported regression results also show that the regression coefficient of the LEV variable is insignificant (p-value = 0.809).Hence, the regreshsion results do not provide support for H3.The results in

SUMMARY AND CONCLUSIONS
This study employs data from the Kuwaiti emerging market to investigate of the association between the amount of CSRD and a number of firm characteristics.In particular, this study aims at providing empirical evidence about factors influential in determining the amount of CSRD made by Kuwaiti firms.Consistent with to findings documented in several prior CSRD studies conducted in different markets (e.g., Bani-Khalid et al., 2017; Muttakin et al., 2015;Alturki, 2014), the results reported in this study reveal a significant association between the amount of CSRD made by Kuwaiti firms and CFP.This indicates that firms with more CSRD are generally firms that are doing well financially.Moreover, the current study's results report a significant association between the amount of CSRD and firm size.This result is similar to findings reported in other related studies (e.g., Bani-Khalid et al., 2017; Alturki, 2014; Macarulla & Talalweh, 2012), and indicates that larger firms, which are expected to be more vulnerable to public scrutiny, tend to disclose greater amounts of CSRD than smaller firms.The results, however, show that the amount of CSRD is not statistically related to firm's leverage nor firm's industry affiliation.Such results are not surprising, though, as similar results are reported in several prior CSRD studies (e.g., Alturki, 2014).The substantiated empirical findings offered by the current study about the association between the amount of CSRD and certain corporate characteristics is expected to be useful to both investors and firm's management in understanding how CSRD are related to these firm characteristics.The results reported by the current study can also be useful to regulator and policy makers in the development of regulations related to CSRD practices by firms in Kuwait.
This study is subject to some worthy of noting limitations.For example, the current study's examination of the amount of CSRD is limited to information disclosed in firms' annual reports, and hence does not include disclosures made though other media and news channels, such as newspapers and television.Therefore, future similar research can be designed to include these other media and news channels in the analysis of the amount of CSRD.Another limitation of the current study is that its examination of CSRD is restricted to firms listed on KSE.Future research, therefore, can be conducted to examine CSRD made by other kinds of firms in the Kuwaiti market.The check-list used to measure the amount of CSRD used in the current study is a modified version of checklists designed by researchers examining CSRD in markets with settings that may be different than that of the Kuwaiti market, and, hence, may not have appropriately captured CSRD made by the Kuwaiti firms.It seems, therefore, important for future research to pay more attention to the use of CSRD checklists that are appropriate to the market settings where the research examination is being conducted.Moreover, other new variables can be tested in the current study such as corporate governance mechanisms.Another valuable avenue for future research is to replicate this kind of investigation using larger and more recent data sets and different research approaches.

Table 1 .
List of some previous CSRD and CFP studies (Muttakin & Khan, 2014;Muttakin et al., 2015;Bani-Khalid et al., 2017)and firm size has been examined empirically in most of previous CSRD research.As indicated earlier, the common idea is that larger firms are expected to disclose more CSRD as a result of the greater pressures they face by the public to report information about their CSRD activities(Cowen et al., 1987).Empirical findings of previous related studies in general suggested a positive association between the amount of firm's CSRD and firm size(Muttakin & Khan, 2014;Muttakin et al., 2015;Bani-Khalid et al., 2017).In light of prior research, larger firms are expected to provide more CSRD than smaller firms, and the second hypothesis of the present study, therefore, is: Investment Management and Financial Innovations, Volume 16, Issue 1, 2019 (1)p://dx.doi.org/10.21511/imfi.16(1).2019.01

Table 2 .
Characteristics of the study sample

Table 5 .
The model equation is as follows:

Table 4 .
Correlation matrix of variables