Final Thesis

Final Thesis

2018 – 2019

Relation between corporate social responsibility (CSR) and firm performance: The moderating role of internationalization and stake holder orientation.

Abstract

Using a sample of 4,307 firms from worldwide, and 28,405 observations over the 2002-2018 period, this study investigates the relationship between CSR initiatives and firms’ financial performance and moderation effects of internationalization and stakeholder orientation on the relationship. This study finds strong evidence that CSRs improve firms’ financial performance. The stakeholder orientation magnifies the positive relationship while the internationalization has limited influence. This study differs from the previous research in two ways: (1) this study provides a global analysis while the majority of the earlier studies only focuses on single region analysis. (2) This study treats internationalization and stakeholder orientation as moderators, which change the relationship between CSRs and financial performance instead of causing better financial performance directly.

Keywords: Corporate social responsibility, financial performance, internationalization, stakeholder orientation

Study Programme:

MSc IFM

1. Introduction

The inclusion of corporate governance issues in financial studies have grown considerably in the past few decades. To meet the growing information demand from different parties of interest, corporate social responsibility (CSR) reports have gradually became a common business practice. With the growing level of attention paid to CSR related issues, the relationship between CSR and firm performance has been a popular topic of debate for over four decades; however, thus far, a consensus has not been reached. Inspired by the positive correlation between CSR and internationalization found by Attig et al. (2016), internationalization will be treated as a catalyst that magnifies the influence of CSR initiatives in this study. Moreover, this study enriches most of the previous studies by addressing the issue proposed by Ioannou and Serafeim (2012). In particular, Ioannou and Serafeim (2012) state that majority of the studies only pay attention to single country or region analysis, while the influences of country-level factors such as institutional environment are overlooked. This study addresses this issue by investigating the impacts of CSR awareness and related institutional developments in different countries based on the relationship between CSR and firm performance.

regions or countries (Hitt et al., 2007, p. 251). The expansion allows firms to have more accessions to competitive advantages in compare with domestic firms. The competitive advantages are represented in the form of increased firm value through enhanced economies of scale and scope (Kogut, 1985), growth opportunities (Porter, 1990), access to new knowledge, production capabilities (Hitt et al., 1997), and so on.

Different from the competitive advantages gained through market expansion, the benefits of CSR disclosure derive from legitimacy development. Specifically, CSR initiatives facilitate communications and information exchange between the firms and the public. By satisfying the people with the information they require, CSR initiatives improve the relationship between firms and business partners, governments, and authoritative institutes. Therefore, CSR initiatives help international companies to develop social networks and communication abilities (Hah & Freeman, 2014). The experience enables firms to deal with the intensified pressures of geographically diversified challenges and help firms to develop legitimacy in host countries (Yang and Rivers, 2009; Jamali, 2010). In general, market expansion unlocks different types of competitive advantages, and CSR initiatives build acceptance among the public which increases the market share, benefits from both perspectives interact with each other and potentially exerts a positive influence on firm financial performance.

awareness and related institutional development—the higher the stakeholder orientation score, the higher the CSR awareness and institutional development. It follows logically that CSR initiatives are valued more in countries with higher CSR awareness than those countries with lower CSR awareness. Furthermore, firms require crucial resources from stakeholders, as these crucial resources enable firms to maintain or gain competitive advantages. When factors outside firms negatively affect them, the support from stakeholders helps firms to recover faster from the consequences of negative business environments (Cheung et al., 2016). Therefore, stronger stakeholder orientation is expected to intensify the influence of CSR initiatives on the financial performances of firms. According to the discussions, the research question of this study is:

How do CSR initiatives affect firm performance under the influence of internationalization and CSR related institutional developments?

The following subsections are organized as follows. The first section describes the theories presented by previous literature and the hypothesis derived from such reasoning. The second section presents the collected data with descriptive statistics and variable measurements. The third section focuses on the econometric models. Finally, the study presents the main empirical results in the fourth section followed by the conclusion and managerial implications of the study.

2. Literature Review & Hypothesis

CSR Initiatives and Firm Performance

managers try to maximize their own benefits at the expenses of the shareholders (Jensen and Meckling, 1976). Within the context of CSR activities, conflicts exist in CSR investments. Cheng et al. 2014 point out that firms normally deviate from their optimal investment because of imperfect market conditions (Cheng et al. 2014), which leads to overinvestment and underinvestment. To be more specific, managers tend to overinvest in CSR activities to increase their personal reputations and to show the public that they pay close attention to social issues. The costs of building the managers’ personal images are borne by the investors because managers sacrifice optimal investment opportunities to pursue their personal interests. As such, agency problems lead to suboptimal performance due to inferior investment choices and destroy the firm’s value. There are many empirical studies that have found evidence to support the incentive for managers to overinvest. For instance, Werbel and Carter (2002) found that engagement in charitable activities is positively correlated to the membership of CEOs in charitable organizations by testing a sample of 160 corporations. Similarly, Barnea and Rubin (2010) found that corporate insiders tend to invest more than the optimal amount of money in CSR as the burdens are borne by the investors and the insiders enjoy the benefits of enhanced personal reputations.

reasoning, a negative relation between CSR initiatives and firm performance due to inefficiency caused by underinvestment has been confirmed in empirical studies (Blanchard et al., 1994; Lang et al., 1991).

On the other hand, stakeholder theory holds that the firm value is influenced by all the parties of interest. These parties include both groups of managers and stakeholders, while the stakeholders are further classified into investing and non-investing stakeholders (shareholders, suppliers, customers, etc.). CSR initiatives are expected to enhance the firm’s performance through balancing the interests of different parties to reduce the risk of resource shortage. In particular, CSR initiatives are used to mitigate information asymmetry between different parties of interest (Freeman, 1984). During the information sharing process, CSR is treated as a mechanism to achieve better information exchanges between insiders and outsiders. Therefore, the effort that firms invest in improvements to CSR related disclosures allow investors to perceive less risk in their investments, which in turn, encourages investors to provide funds under more lenient conditions and reduce the cost of capital for firms (Cui et al., 2018). To support this point of view, Ruf, Muralidhar, Brown, Janney, and Paul (2001) found that the CSR engagements of corporations are positively associated with the financial performances of firms.

reputation. Ben Brik, Rettab, and Mellahi (2011) found that CSR and market orientation creates a synergy, which is positively correlated to business performance. Similarly, Attig et al. (2013) found that CSR increases the credit ratings of firms, which leads to lower capital costs (El Ghoul, Guedhami, Kim, & Park, 2018) and market frictions (Attig, Cleary, El Ghoul, & Guedhami, 2014). Taken together, drawing upon the stakeholder theory and resource-based view, CSR initiatives lead to better firm performance.

In conclusion, the findings regarding the correlation between firm performance and CSR initiatives remain mixed; from the perspective of the agency theory, the correlation is expected to be negative, whereas from the perspective of the stakeholder theory and resource-based view, the correlation is expected to be positive. Hence, the actual effect of CSR initiatives on firm performance is open for discussion in this study. Given the above discussions, the first hypothesis has been formulated as follows:

Hypothesis H1: CSR initiatives have no influence on the firm performance

Moderating Effect of Internationalization

expanding to other markets is a strategy designed to seek more opportunities in foreign markets, which potentially benefit firms. Moreover, the disadvantages encountered by firms are partially solved by developing social ties, which consist of personal networks (Ellis, 2010), interpersonal relationships (Harris & Wheeler, 2005), social networks (Ellis, 2010), and relational networks (Chen & Chen, 1998). By investing in the development of social ties, firms enhance the trust and acceptance they receive from their partners; the trust between the firms and their partners promotes better relationships, which in turn, reduces the transaction cost (Ellis, 2010) and uncertainty (Sharma & Blomstermo, 2003) during the internationalization.

With respect to the development of social ties, firms have the incentive to invest in social issues. Firms face stakeholders with various demands, norms, and complexities in various countries. Meanwhile, firms also work in regions with diverse economic, political, and institutional backgrounds (Sambharya, 1996). Such diversity poses challenges to firms and encourages firms to commit to CSR principles and invest in CSR disclosures. As such, firms overcome the pressure from their stakeholders, which further promotes the development of social ties. As a matter of fact, the benefit of CSR disclosure is multiplied by the decision to internationalize, since expanding to foreign markets enables MNEs to enhance their reputations in more markets instead of merely relying on domestic markets (Lange et al., 2011). According to the empirical study conducted by Thams et al. (2016), stakeholders have better impressions of MNEs after they conquer initial barriers in the internationalization process. Given the above facts, the second hypothesis has been formulated as follows:

Moderating Effect of Stakeholder Orientation

Most of the literature in the same domain conduct research based on one specific country, while country-level factors exert influence over CSR related issues. As was pointed out by Dhaliwal et al. (2014), stakeholder orientation affects firms’ access to crucial resources; stakeholders in countries with stronger stakeholder orientation levels have greater influences on firm operations (Chen, 2009). Drawing on the stakeholder theory, MNEs with better CSR disclosures will show superior performance in countries with greater stakeholder orientation. Because CSR initiatives provide more information required by stakeholders to evaluate firms, and more information included in financial reports is likely to increase the quality of reports, as an outcome, the risk perceived by stakeholders decreases. With less risk perceived by the stakeholders, firms raise money more easily under more lenient conditions, which allows the firm to sustain competitive advantages or recover faster from negative economic situations (Cheung et al., 2016). This is the case, especially when stakeholders are in a country with a higher stakeholder orientation score, because stakeholder orientation measures CSR awareness. Therefore, firms that devote more effort to CSR initiatives provide more valuable information to stakeholders to maintain their competitive advantages. This is empirically supported by Dhaliwal et al.’s (2011) study. Following this vein, CSR initiatives are more appreciated by stakeholders in countries with stronger stakeholder orientation than those with lower stakeholder orientation. Thus, the last hypothesis is formulated:

3. Data and Research Design

3.1. Sample Selection

The data covers firms from 31 countries, the research time length during the analysis lasts from year 2002 to 2018. The data source for the dependent variables and firm-level independent variables were obtained from DataStream. For much of the financial study, the financial firms were eliminated from the dataset because their characteristics distinguish them from regular companies. In particular, because of the different business model applied to financial firms, the leverage levels of financial firms are higher than that of other companies—a high level of leverage is a sign of financial distress in non-financial firms while it has a different meaning for financial firms (Fama & French, 1992).

The sample selection process was started by constructing a list of ISIN codes from ASSET4, which included all firms with ESG scores. Later, the SIC codes were matched to the ISIN codes to identify the industry of each company. As such, financial institutes (SIC code range from 6000–6999) were identified and erased from the dataset. After deleting the financial companies, the companies remaining in the sample were those with financial information from the period of 2002 to 2018. Additionally, all the financial data were unified in US dollars. During the last phase, ratios used for the regression analysis were calculated.

strategy. Evaluation of these dimensions was meant to capture the performance of different CSR related practices, and firms with higher ESG scores were deemed to be more stakeholder oriented (Cheung et al., 2018).

In respect to the country-level variables, the sample consists of companies in 31 countries, which are presented in Table 1. Country-level variables related to macroeconomics were retrieved from World Bank, and the data was available from year 2002 to 2017. Lastly, the stakeholder orientation index was acquired from previous research provided by Dhaliwah et al. (2012). According to the previous literature, there are 31 stakeholder orientation scores available for each of the countries in the sample. Stakeholder orientation was constructed based on the ranking information from different law related domains, which are similar to creditor rights. Cho et al. (2014) explain that this is because laws rarely undergo dramatic changes from year to year, so the ranking in general is consistent; therefore, the stakeholder orientation score remained constant throughout the research period.

Table 1. Sample Distribution by Country

Country Obs. Percentage Country Obs. Percentage Country Obs. Percentage

Austria 143 0.50% France 952 3.30% Norway 190 0.66%

Australia 1,924 6.68% United Kingdom 2,907 10.09% New Zealand 158 0.55%

Belgium 196 0.68% Greece 128 0.44% Philippine 81 0.28%

Brazil 443 1.54% Hong Kong 356 1.24% Portugal 89 0.31%

Canada 1,886 6.55% India 546 1.90% Sweden 522 1.81%

Switzerland 636 2.21% Italy 344 1.19% Singapore 343 1.19%

Chile 117 0.41% Japan 3,744 13.00% Thailand 132 0.46%

Germany 846 2.94% Korea 574 1.99% United States 9,170 31.83%

Denmark 259 0.90% Mexico 159 0.55% South Africa 549 1.91%

Spain 396 1.37% Malaysia 247 0.86%

Finland 320 1.11% Netherlands 448 1.56%

3.2. Measurements

Dependent Variable

Firm Performance (ROA/ROE/Tobin’s Q): The dependent variable in this study is firm

performance, this study adopts return on assets (ROA) as a proxy for the firm’s financial performance (e.g., Miller et al., 2018). The measurement captures the ability of firms to generate revenue with each unit of asset. Moreover, two alternative measurements were used in the robustness test: like return on assets, return on equities measures the ability of firms to generate returns from each unit of equity. Another is Tobin’s Q, which indicates whether a given business is overvalued or undervalued.

Main Independent Variables

CSR Initiatives (ESG): One of the main independent variables is CSR initiatives, and it is

proxied by ESG scores. ESG scores consist of three areas, which are environment, social, and governance. These measurements capture the firms’ responses to various social issues and the stakeholders’ interests (Attig et al., 2016), with scores ranging from 0 to 100—the lower the score is, the lower the CSR awareness is in a country.

Internationalization (F_Sales): The level of internationalization is another firm-level variable.

to expand sales abroad; therefore, foreign sales are the measurement of this variable. The calculation is the total amount of foreign sales divided by the total sales.

Stakeholder Orientation (Stake): The level of stakeholder orientation is a country-level

variable, and the data was retrieved from Dhaliwal et al. (2012). Stakeholder orientation consists of four law related proxies.

The first proxy is stake law, which is the mean value of indices that measure the stringency of employment, social security, collective bargaining, and human rights laws. Stake law primarily evaluates a country’s protection of labor rights and is the average ranking score of four law indices: (1) employment laws; (2) social security laws; (3) collective relation laws; and (4) human rights laws. In general, a higher value in stake law implies a higher level of stakeholder orientation.

The second proxy is CSR-related disclosure laws, which classifies companies in a country into two categories of commercial and pension fund firms. The proxy has a value of one if a country mandates either commercial or pension fund firms to publish CSR information; the proxy has a value of two if a country mandates both commercial and pension fund firms disclose CSR information; otherwise, the proxy takes a value of zero. Therefore, higher values for CSR related disclosure laws also contribute to higher levels of stakeholder orientation.

and non-commercial organizations scaled by the multiple of the millions of people in a country. Thus, a higher stakeholder orientation is expected if the ranking score of CSR awareness in a country is higher.

The last proxy is an alternative measurement of CSR awareness, which is obtained through a survey that directly captures the attitudes of the managers, inventors, and public towards social issues. The proxy is the mean ranking score of four indices: (1) sustainable development priority; (2) ethical practice implementation; (3) social responsibility of business leader; and (4) corporate responsibility competitiveness index. Higher CSR awareness values among stakeholders means higher levels of stakeholder orientation.

Firm-Level Control Variables

The following methods have been adopted by most literature in the financial domain. The firm-level control variables, such as profitability (Pro), leverage (Lev), capital expenditure (CapEx), market to book ratio (M2B), intangibility (I_Asset), and firm size (Size), were included in the regression analysis to mitigate the influence of firm characteristics.

Profitability (Pro) is the earnings before interest and tax ratio scaled by total sales, while the level of leverage (Lev) is the total amount of debt divided by the total assets. Both profitability and leverage level are assumed to capture the unused resources in companies, as there has been evidence to suggest that firms with more unused resources are more willing to invest in CSR practices (Waddock and Graves, 1997).

outstanding multiplied by the share prices minus the book value of the equity plus the book value of the asset. The reason for using the logarithm value is to mitigate the influence of companies with extremely high value market capitalizations; thus, the observations are more comparable. The purpose for adding the market to book ratio into the regression model is to isolate the part of variation caused by intangible assets, because a higher value demonstrates a higher growth potential (Kang, 2013).

Intangibility (I_Asset) is the total intangible assets to total assets; this variable is assumed to capture the characteristic that firms with more intangible assets generate higher returns than firms with low intangible assets. Capital expenditure (CapEx) is the total spent on acquiring equipment, property, and plant divided by the total sales. Capital expenditure indicates the cost efficiency of a firm.

Lastly, firm size (Size) is the logarithm of the market firm value. This control variable captures the extent to which firms are exposed to the outside world. Given the fact that large firms are more influential on the public, these firms receive more attention from the public, which causes these large firms to face more pressure to behave in a socially responsible way (Brammer et al., 2009).

Country-Level Control Variables

According to prior studies, this study includes country-level variables (Attig et al., 2016) which may influence the firm performance. The country level variables include GDP growth (GDP_G), GDP per capital (GDP_C), and inflation (Inflation). The data was available from

World Bank. In particular, these country-level control variables have a potential influence on

Statistical Summary and Correlation Matrix

Table 2 presents the statistical summary of all variables—for each of the dependent and firm-level independent variables—and the number of firm-year observations is 28,405. The number of observations for the country-level variables is 447.

The table shows the dependent variable first. In the table, it can be seen that the Tobin’s’ Q ranges from 0.26 to 29.46, while the median and mean are 1.44 and 1.84, respectively, which suggests that most of the values tend to be small and centered around 1.84, even though the highest value is 29.46. The median of ROE is 0.18 and the variable ranges from -46.18 to 107.13, which suggests that negative values are much more frequent than positive values.

Table 2. Summary Statistic

Obs. Mean Median Min Max S.D.

Dependent Variables Tobins_ Q 28,405 1.84 1.44 0.26 29.46 1.31 ROE 28,405 0.21 0.18 -46.18 107.13 1.01 ROA 28,405 0.08 0.08 -2.41 1.48 0.11 Firm-level variables ESG 28,405 56.61 61.52 2.42 98.71 29.87 M2B 28,405 1.06 1.05 0.76 2.30 0.06 Pro 28,405 0.12 0.10 -1.00 1.00 0.16 CapEx 28,405 0.05 0.04 0.00 0.93 0.05 F_Sales 28,405 0.40 0.37 0.00 34.64 0.39 Size 28,405 15.40 15.36 8.90 20.50 1.45 Lev 28,405 0.54 0.55 0.01 1.00 0.19 I_Asset 28,405 0.20 0.12 0.00 0.96 0.20 Country-level variables GDP_G 447 0.02 0.02 -0.09 0.15 0.02 GDP_C 447 10.55 10.72 6.90 11.54 0.66 Inflation 447 0.02 0.02 -0.06 0.14 0.02 Stake 31 -0.40 -0.95 -2.73 2.95 1.33

Note: This table reports the summary statistic for each variable. The sample size of firm-level variables is 28,405, representing 4,307

unique firms across 31 countries from 2002 to 2018, each of the country-level variables is composed of 447 unique values. Tobins Q is the sum of market capitalization and total liabilities scaled by the sum of common stock equities and total liabilities; ROE is earnings before interests and taxes (EBIT) to total equities; ROA is EBIT to total assets; ESG is the proxy scores for CSR initiatives obtained from ASSET4; M2B is the market capitalization divided by common stock equity; Pro, CapEx, and F_Sales are Pro, Capital expenditure, and foreign sales to total sales respectively; Size is a logarithm of total assets; Lev is a ratio of total liabilities to total assets;

I_Asset is the intangible assets to total assets; GDP_G is defined as the GDP growth rate; GDP_C is defined as a logarithm of GDP per

The second part of the table presents the firm-level independent variables. The variable Pro is a denotation of profitability, which is earnings before interest and taxes scaled by total sales. Table 2 only presents the summary of the statistics after excluding extreme values based on the 1.5 interquartile rule. The absolute maximum and minimum values of the variables are close to 1 with a median and mean close to 0. F_Sales represents the ratio of foreign sales to total sales—like Tobin’s Q, it has a large maximum value of 34.64 while most of the values are centered around 0.4.

The third part of the table shows the country-level variables. Stake is a denotation of stakeholder orientation. There are only 31 observations, which match the number of countries. Because stakeholder orientation is a law related indicator, it rarely changes throughout the years.

Table 3. Correlation Matrix

Tobins_ Q ROE ROA ESG M2B Pro CapEx F_Sales Size Lev I_Asset GDP_G GDP_C Inflation Stake

Tobins_ Q 1.00 ROE 0.12 1.00 ROA 0.45 0.32 1.00 ESG -0.08 0.05 0.06 1.00 M2B 0.65 0.23 0.42 -0.04 1.00 Pro 0.18 0.11 0.38 0.04 0.20 1.00 CapEx -0.03 -0.01 -0.04 -0.03 0.00 0.03 1.00 F_Sales 0.01 -0.02 0.00 0.21 -0.01 -0.02 -0.05 1.00 Size -0.29 0.01 -0.07 0.47 -0.26 0.06 -0.05 0.08 1.00 Lev -0.18 0.10 -0.08 0.19 0.17 -0.16 -0.04 -0.04 0.34 1.00 I_Asset 0.04 0.01 0.00 0.03 0.13 0.03 -0.30 0.06 -0.01 0.08 1.00 GDP_G 0.09 0.02 0.05 -0.05 0.13 0.12 0.05 -0.03 -0.04 -0.04 -0.04 1.00 GDP_C -0.04 -0.01 -0.09 -0.02 0.01 -0.10 -0.05 0.07 -0.02 0.00 0.15 -0.29 1.00 Inflation 0.07 0.03 0.10 0.06 0.09 0.13 0.09 -0.04 -0.04 0.02 0.05 0.27 -0.47 1.00 Stake -0.08 -0.02 -0.05 0.13 -0.07 -0.04 0.05 0.26 -0.14 -0.01 0.10 -0.10 0.35 -0.12 1.00

Note: This table reports the correlation matrix for the regression variables. The sample consists of 29,405 firm-year observations, representing 4,307

unique firms across 31 countries from 2002 to 2018. Tobins Q is the sum of market capitalization and total liabilities scaled by the sum of common stock equities and total liabilities; ROE is earnings before interests and taxes (EBIT) to total equities; ROA is EBIT to total assets; ESG is the proxy scores for CSR initiatives obtained from ASSET4; M2B is the market capitalization divided by common stock equity; Pro, CapEx, and F_Sales are EBIT, Capital expenditure, and foreign sales to total sales respectively; Size is a logarithm of total assets; Lev is a ratio of total liabilities to total assets;

I_Asset is the intangible assets to total assets; GDP_G is defined as the GDP growth rate; GDP_C is defined as a logarithm of GDP per capita; lastly, Inflation is the inflation ratio. The correlation coefficient that are reported in boldface are significant at 5% level.

4. Regression Models

The regression model is shown in equation (1), where the ROA is a denotation of the return on asset of firm i, in industry j, in country k, in year t. The main independent variables include

ESG, F_Sales, and Stake. ESG is the value of the ESG scores, F_Sales is the proxy for the level

of internationalization, and Stake represents the proxy for stakeholder orientation. Firm

Controls are a set of control variables of the firm characteristics and Country Controls are a set

of the country-level control variables. The F.E. is a denotation of the fixed effects, which consists of a set of industry dummies and year dummies. To control for the endogeneity problem, the independent variables are lagged values, and a set of the year and industry dummies are included in regression analysis.

𝑅𝑂𝐴_{𝑖,𝑗,𝑘,𝑡} = 𝛽₀+ 𝛽₁∗ 𝐸𝑆𝐺_{𝑖,𝑗,𝑘,𝑡−1}+ 𝛽₂∗ 𝐹_𝑆𝑎𝑙𝑒𝑠_{𝑖,𝑗,𝑘,𝑡−1}+ 𝛽₃∗ 𝑆𝑡𝑎𝑘𝑒_{𝑘,𝑡−1} + 𝛽₄∗ 𝐸𝑆𝐺 ∗ 𝐹_𝑆𝑎𝑙𝑒𝑠_{𝑖,𝑗,𝑘,𝑡−1}+ 𝛽₅∗ 𝐸𝑆𝐺 ∗ 𝑆𝑡𝑎𝑘𝑒_{𝑖,𝑗,𝑘,𝑡−1} + ∑ 𝛽_𝑚∗ 𝐹𝑖𝑟𝑚 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠_{𝑖,𝑗,𝑘,𝑡−1}

+ ∑ 𝛽_𝑛 ∗ 𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠_{𝑖,𝑗,𝑘,𝑡−1}+ 𝐹. 𝐸. +𝜀

The main purpose of the analysis was to test the relationship between firm performance and CSR initiatives with the moderating effect of internationalization and CSR awareness. Therefore, focus was placed on the coefficient of the ESG scores and two interaction terms. The regression analysis proceeded as follow: The analysis began with test the relation between the independent variables without any extreme values. Then, a robustness test was conducted. Instead of the ROA, the robustness test checked if the three hypotheses still held after using the measurement of the firm performance. Finally, the sensitivity of the results to extreme values are tested by replicating the former analysis.

To test the first hypothesis, the regression model did not include the interaction terms, the dependent variable was set to the ROA, and the main independent variable was ESG. After making the inference about the relation between firm performance and CSR initiatives, the study moved onto the second hypothesis test. To test the second hypothesis, the interaction terms of foreign sales and ESG scores were added into the model to assess the moderating effect of internationalization. Lastly, the third hypothesis test was conducted by adding the interaction terms of stakeholder orientation and CSR initiatives into the first hypothesis model so that the influence of CSR awareness on the original relationship could be discovered. The robustness test followed the same process; however, the alternative dependent variables were adopted instead of ROA.

5. Empirical Result

Main Evidence

the regression analysis did not suggest any causal relation. In Model (1), the firm performance proxy (ROA) was regressed on the main explanatory variable, which is the firms’ CSR performance (ESG) and a set of control variables. The empirical results were consistent with the stakeholder theory: there is a statistically significant correlation between ROA and ESG scores, and the result is significant at the 1% (t = 2.20) level. This result is in line with Cui et al.’s (2018) finding that CSR serves as a mechanism, which facilitates communication between firms and stakeholders. Therefore, information asymmetry is reduced during the information exchange process, which leads to a lower perception of risk by investors. With lower perceived risk, investors are more willing to provide funds at lower cost. Alternatively, from the resource-based view, CSR disclosure is treated as a strategy to build the firm’s reputation (McWilliams and Siegel, 2011). Although the result is statistically significant, the coefficient is not economically significant. As shown in the table, the coefficient is close to 0. In particular, with each unit increase in the ESG score, there is only a 0.03% increase in the returns. The small magnitude of the coefficient potentially indicates a large effort devoted to the CSR performance contributing to small increases in returns.

competitive advantages; the negative coefficient indicates that international firms suffer more obstacles than benefits. Taken together, the results indicate that CSR initiatives help firms to gain legitimacy in different markets during the internationalization process. In other words, firms that sell their products and/or services to foreign markets perform better when they enhance their CSR disclosures. This result provides further support to the stakeholder theory and resource-based view, as CSR is a strategy, which is designed to build social networks in foreign markets to bring competitive advantages presented in the form of higher returns.

Model (3) shows the results of the moderating effect of CSR awareness on firm performance. In addition to Models (1) and (2), Model (3) includes several country-level variables. The significance and coefficient magnitude are low, which is similar to those of the other two models. Stakeholder orientation (Stake) is not significantly correlated to firm performance, which means that CSR awareness alone does not exert an influence on the performances of the firms in the 31 countries. In respect to the interaction term, it is significant at the 1% level (t = 5.00), with a coefficient equal to 0.0001. As such, the inference is that firms are more successful in countries with higher CSR awareness when they have CSR disclosure related policies. The positive relationship between the interaction term and ROA is in line with the hypothesis based on Cheung et al.’s (2016) empirical results, which suggests that firms that devote much effort to CSR activities earn trust, support, and crucial resources from stakeholders in countries with high CSR awareness. The crucial resources held by stakeholder allow firms to overcome bad consequences from negative business environments and/or perform better than their peer competitors.

internationalization process and in countries with high CSR awareness. Therefore, the evidence tends to lean towards supporting the stakeholder theory and resource based view. In particular, firms build their trust and reputations among stakeholders through disclosing CSR related information and reducing information asymmetry. Therefore, CSR engagement is a valuable strategy that enables firms to assess the sustainable competitive advantages to be gained from all parties of interests.

Table 4. Regression Results

Dependent variable ROA

(1) (2) (3) Firm-level variables ESG 0.0003*** 0.0002*** 0.0002*** (13.80) (8.56) (10.32) M2B 0.6572*** 0.6569*** 0.6704*** (66.66) (66.62) (66.67) Pro 0.2018*** 0.2019*** 0.1979*** (55.39) (55.4) (54.05) CapEx -0.0648*** -0.0655*** -0.0679*** (-5.71) (-5.77) (-5.97) Size 0.0011** 0.0009* 0.0019*** (2.20) (1.95) (3.67) Lev -0.0743*** -0.0742*** -0.0774*** (-21.92) (-21.90) (-22.78) I_Asset -0.0353*** -0.0353*** -0.0332*** (-12.41) (-12.4) (-11.37) F_Sales -0.0119*** (-3.45) Country-level variables Stake 0.0027 (-1.45) GDP_G -0.2613*** (-7.04) GDP_C -0.0108*** (-10.24) Inflation 0.1133*** (3.25) Interaction terms ESG  F_Sales 0.0001** (2.20) ESG  Stake 0.0001*** (5.00)

Firm fixed effect Yes Yes Yes

Year fixed effect Yes Yes Yes

Number of observations 28,405 28,405 28,405

𝐴𝑑𝑗 − 𝑅2 _{0.3221 0.3224 0.3272}

Note: This table reports the estimated result ordinary least square (OLS) regression, the numbers presented below the coefficient in

Regression Results with Extreme Values

During the data screening process, 397 firm-year observations were excluded from the sample due to the extremely large negative values. However, this is the case when firms generate small amounts of money relative to their total expenses, so the profitability ratios could be extremely negative. In this section, the negative values were added back into the total sample. Then, this study replicated the same procedure for the regression analysis conducted above to test the sensitivity of the empirical results to extreme values.

Although the test produced different results than that of previous test after including the extreme values, considering that most of the differences were caused by observations in the mining industry, which may be due to specific industry characteristics and not representative of the whole sample, this study will stick to the results after excluding the extreme values.

Robustness Test

In this section, the regression result of the alternative measurements (ROE and Tobin’s Q) for firm performance will be focused on. As exhibited in Table A. 2. in the appendix, when the dependent variable is ROE, the ESG scores are significantly correlated to the firm performance across all models. So, the significant coefficients provide more evidence to confirm the first hypothesis. However, although the interaction terms of internationalization level and CSR initiatives are positively associated with firm performance, the coefficient is neither statistically nor economically significant. This evidence implies that CSR disclosures do not bring firms benefits when they expand their services and/or products into foreign markets. The significant positive coefficient of the interaction term in Model (3) confirms the third hypothesis.

In conclusion, the evidence produced by Models (1) and (3) are consistent with the previous empirical test conducted in this study when measurement of the firm’s financial performance was ROA. Specifically, the positive moderation effect of CSR awareness has been identified in all models using alternative firm performance proxies. Additionally, CSR initiatives improve the firms’ financial performances in general. In respect to Model (2), given the fact that the coefficient of the interaction terms in Model (2) are all positive, but only statistically significant in the regression test when ROE is the dependent variable, it is concluded that hypotheses two is partially supported.

6. Conclusion

This study investigated the moderation effects of the internationalization level and CSR awareness on the relationship between CSR initiatives and firm financial performance.

Since the aforementioned tests did not include samples with extremely low profitability and considering that huge losses are still a case in reality, this paper also included a test with extreme values in the data. In line with the previous results, CSR initiatives exerts a positive influence on firm financial performance, but all interaction terms are statistically insignificant. After investigating the composition of the observations with huge amount of losses, most of the observations were from the mining industry, which are not considered as representative of the whole sample. Due to the lack of representativeness, this study chose to stick to the previous results.

In summary, the empirical results confirm the three hypotheses proposed by this paper, which indicates the importance of stakeholders and/or competitive advantages to firms. The implication for firms is to respond to the demands of stakeholders and to treat CSR initiatives as a crucial source of competitive advantages. In the meantime, it should be noted that the economic significance of ESG scores are low, which implies that increases in financial performance requires much more effort devoted to CSR development. Managers should make a trade-off between the benefits of improving CSR related activities and the costs associated with it.

countries in the sample are developed countries and accounted for more than 90% of the observations. For the reason that institutional architecture differs and are still evolving in emerging markets compared to developed countries, future studies can extend this research by adding more developing countries into the empirical analysis.

Another limitation was related to the channel through which how firms realize benefits. Although this study used the stakeholder theory and resource-based view to support the fact that CSR initiatives lead to better firm performance, but this study was unable to distinguish the exact source of benefits and the channel through which firms achieve it. Future studies can therefore delve into the approach as to how firms achieve better performances through CSR initiatives.

7. Managerial Implications

Appendix A.

Table A. 1. Regression Results: Including Extreme values

Dependent variable ROA

(1) (2) (3) Firm-level variables ESG 0.0003*** 0.0003*** 0.0003*** (13.59) (11.87) (10.06) M2B 0.7990*** 0.7992*** 0.8066*** (74.94) (74.95) (74.14) Pro 0.0081*** 0.0082*** 0.0079*** (31.16) (31.27) (30.69) CapEx -0.0651*** -0.0644*** -0.0722*** (-5.19) (-5.14) (-5.77) Size 0.0082*** 0.0082*** 0.0092*** (15.18) (15.17) (15.95) Lev -0.1169*** -0.1169*** -0.1186*** (-31.67) (-31.67) (-32.12) I_Asset -0.0257*** -0.0257*** -0.0223*** (-7.87) (-7.87) (-6.67) F_Sales 0.0032** (2.46) Country-level variables Stake 0.0030*** (2.75) GDP_G -0.2002*** (-4.70) GDP_C -0.0166*** (-13.74) Inflation -0.0223*** (-6.67) Interaction terms ESG  F_Sales 0.0000 (-1.16) ESG  Stake 0.0000 (-0.55)

Firm fixed effect Yes Yes Yes

Year fixed effect Yes Yes Yes

Number of observations 28,802 28,802 28,802

𝐴𝑑𝑗 − 𝑅2 _{0.3221 0.3224 0.3272}

Note: This table reports the estimated result ordinary least square (OLS) regression after including observations with extreme values of

Table A. 2. Regression Result: Robustness Test

Dependent variables ROE Tobins_Q

(1) (2) (3) (1) (2) (3) Firm-level variables 0.0008*** 0.0009*** 0.0005** 0.0006*** 0.0004 0.0007*** ESG (3.79) (3.24) (2.29) (3.11) (1.59) (3.39) 3.4816*** 3.4891*** 3.5166*** 14.3171*** 14.3135*** 14.2984*** M2B (32.43) (32.49) (32.02) (143.99) (143.9) (140.78) 0.4916*** 0.4884*** 0.4911*** 0.1183*** 0.1196*** 0.0819** Pro (12.39) (12.31) (12.28) (3.22) (3.25) (2.22) -0.3388*** -0.3556*** -0.3456*** -0.8810*** -0.8778*** -0.9091*** CapEx (-2.74) (-2.88) (-2.78) (-7.7) (-7.66) (-7.92) 0.0207*** 0.0211*** 0.0234*** -0.0079* -0.0085* -0.0095* Size (3.96) (4.03) (4.20) (-1.65) (-1.75) (-1.86) 0.3385*** 0.3344*** 0.3317*** -2.0602*** -2.0587*** -2.0615*** Lev (9.18) (9.06) (8.94) (-60.29) (-60.21) (-60.06) -0.1334*** -0.1279*** -0.1367*** -0.3541*** -0.3560*** -0.3182*** I_Asset (-4.31) (-4.12) (-4.28) (-12.35) (-12.4) (-10.78) -0.0609 0.0347 F_Sales (-1.63) (-0.54) Country-level variables 0.0289*** -0.0195** Stake (2.78) (-2.03) -0.5114 -0.0146 GDP_G (-1.26) (-0.04) -0.0034 -0.0734*** GDP_C (-0.3) (-6.91) 0.06189 0.3897 Inflation (0.16) (1.11) Interaction terms 0.0001 0.0005 ESG  F_Sales (0.21) (1.06) -0.0004*** 0.0002** ESG  Stake (-3.07) (1.98)

Firm fixed effect Yes Yes Yes Yes Yes Yes

Year fixed effect Yes Yes Yes Yes Yes Yes

Number of observations

28,405 28,405 28,405 28,405 28,405 28,405

𝐴𝑑𝑗 − 𝑅2 _{0.0680 0.0680 0.0678 0.5312 0.5312 0.5327}

Note: This table reports the estimated result ordinary least square (OLS) regression by adopting alternative measurements: ROE and

Reference List

Attig, N., Boubakri, N., Ghoul, S., & Guedhami, O. 2016. Firm Internationalization and Corporate Social Responsibility. Journal of Business Ethics, 134(2): 171–197.

Attig, N., Cleary, S. W., El Ghoul, S., & Guedhami, O. (2014). Corporate Legitimacy and Investment – Cash flow sensitivity. Journal of Business Ethics, 121, 297–314.

Aupperle, K. E., Carroll, A. B., & Hatfield, J. D. 1985. An Empirical Examination of The Relationship Between Corporate Social Responsibility and Profitability. Academy of Management Journal, 28(2): 446–463.

Barnea, A. & Rubin, A. 2006. Corporate Social Responsibility as a Conflict Between Shareholders. Journal of

Business Ethics, 97(1): 71–86.

Barney, J. 1991. Firm resources and sustained competitive advantage. Journal of Management, 17(1): 99–120. Bartlett, C. A., & Ghoshal, S. 1989. Managing across borders: the transnational solution. Boston, MA: Harvard Business School Press.

Benlemlih, M., & Bitar, M. 2016. Corporate Social Responsibility and Investment Efficiency. Journal of

Business Ethics, 148(3): 647–671.

Blanchard, O., & Lopez-De-Silane, F. 1994. What do Firms do with Cash Windfalls? Journal of Financial

Economics, 36(3): 337–360.

Bouslah, K., Kryzanowski, L., & Mzali, B. 2013. The Impact of The Dimensions of Social Performance on Firm Risk. Journal of Banking & Finance, 37(4): 1258–1273.

Brik, A. B., Rettab, B., & Mellahi, K. 2010. Market Orientation, Corporate Social Responsibility, and Business Performance. Journal of Business Ethics, 99(3): 307–324.

Buchanan, B., Cao, C. X., & Chen, C. 2018. Corporate Social Responsibility, Firm Value, and Influential Institutional Ownership. Journal of Corporate Finance, 52: 73–95.

Chen, C. X. 2009. Who Really Matters? Revenue Implications of Stakeholder Satisfaction in a Health Insurance Company. The Accounting Review, 84(6): 1781–1804.

Chen, H., & Chen, T.-J. 1998. Network Linkages and Location Choice in Foreign Direct Investment. Journal of

International Business Studies, 29(3): 445–467.

Cheng, B., Ioannou, I., & Serafeim, G. 2014. Corporate Social Responsibility and Access to Finance. Strategic

Management Journal, 35(1): 1–23.

Cheung, Y.-L., Tan, W., & Wang, W. 2016. National Stakeholder Orientation, Corporate Social Responsibility, and Bank Loan Cost. Journal of Business Ethics, 150(2): 505–524.

Cui, J., Jo, H., & Na, H. 2018. Does Corporate Social Responsibility Affect Information Asymmetry? Journal of

Business Ethics, 148(3): 549–572.

Dhaliwal, D. S., Li, O. Z., Tsang, A., & Yang, Y. G. 2011. Voluntary Nonfinancial Disclosure and the Cost of Equity Capital: The Initiation of Corporate Social Responsibility Reporting. The Accounting Review, 86(1): 58– 100.

Dhaliwal, D. S., Li, O. Z., Tsang, A., & Yang, Y. G. 2014. Corporate Social Responsibility Disclosure and the Cost of Equity Capital: The Roles of Stakeholder Orientation and Financial Transparency. Journal of Accounting

and Public Policy, 33(4): 328–355.

Ellis, P. D. 2010. Social ties and international entrepreneurship: Opportunities and constraints affecting firm internationalization. Journal of International Business Studies, 42(1): 99–127.

Fama, E. F., & French, K. R. 1992. The Cross‐Section of Expected Stock Returns. The Journal of Finance, 47(2): 427–465.

Freeman, R. E. (1984). Strategic Management: A Stakeholder Approach. Boston, MA: Pitman.

Hah, K., & Freeman, S. 2013. Multinational Enterprise Subsidiaries and their CSR: A Conceptual Framework of the Management of CSR in Smaller Emerging Economies. Journal of Business Ethics, 122(1): 125–136. Harris, S., & Wheeler, C. 2005. Entrepreneurs Relationships for Internationalization: Functions, Origins and Strategies. International Business Review, 14(2): 187–207.

Hitt, M. A., Hoskisson, R. E., & Kim, H. 1997. International Diversification: Effects on Innovation and Firm Performance in Product-Diversified Firms. Academy of Management Journal, 40(4): 767–798.

Hitt, M. A., Ireland, R. D., & Hoskisson, R. E. (2007). Strategic Management: Competitiveness and Globalization (7th ed.). Mason, OH: South-Western.

Jamali, D. 2008. A Stakeholder Approach to Corporate Social Responsibility: A Fresh Perspective into Theory and Practice. Journal of Business Ethics, 82(1): 213–231.

Jensen, M., & Meckling, W. 1976. Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics, 3(4): 305–360.

Jiao, Y. 2010. Stakeholder Welfare and Firm Value. Journal of Banking & Finance, 34(10): 2549–2561. Jo, H., & Harjoto, M. 2011. Corporate Governance and Firm Value: The Impact of Corporate Social Responsibility. Journal of Business Ethics, 103(3): 351–383.

Kruger, P. 2015. Corporate Goodness and Shareholder Wealth. Journal of Financial Economics, 115(2): 304– 329.

Kogut, B. 1985. Designing global strategies: Profiting from operational flexibility. MIT Sloan Management

Review, 21, 27–38.

Lang, L. H. P., Stulz, R. M., & Walkling, R. A. 1991. A Test of the Free Cash Flow Hypothesis: The Case of Bidder Returns. Journal of Financial Economics, 29(2).

Lange, D., Lee, P. M., & Dai, Y. 2011. Organizational Reputation: A Review. Journal of Management, 37(1): 153–184.

Li, J., Zhou, C., & Zajac, E. J. 2009. Control, collaboration, and productivity in international joint ventures: theory and evidence. Strategic Management Journal, 30(8): 865–884.

McWilliams, A., & Siegel, D. 2001. Corporate Social Responsibility: A Theory of the Firm Perspective. The

Academy of Management Review, 26(1): 117–127.

Miller, S. R., Eden, L., & Li, D. 2018. CSR Reputation and Firm Performance: A Dynamic Approach. Journal

of Business Ethics, 1–8.

Myers, S. 1984. Capital Structure Puzzle. The Journal of Finance, 39(3): 574–592.

Qian, G., Li, L., Li, J., & Qian, Z. 2008. Regional Diversification and Firm Performance. Journal of International

Business Studies, 39(2): 197–214.

Ruf, B. M., Muralidhar, K., Brown, R. M., Janney, J. J., & Paul, K. 2001. An Empirical Investigation of the Relationship Between Change in Corporate Social Performance and Financial Performance: A Stakeholder Theory Perspective. Journal of Business Ethics, 32(2): 143–156.

Sambharya, R. B. 1996. Foreign Experience of Top Management Teams and International Diversification Strategies of U.S. Multinational Corporations. Strategic Management Journal, 17(9): 739–746.

Sharma, D., & Blomstermo, A. 2003. The Internationalization Process of Born Globals: a network view. International Business Review, 12(6): 739–753.

Thams, Y., Alvarado-Vargas, M. J., & Newburry, W. 2016. Geographical diversification as a predictor of MNC reputations in their home nations. Journal of Business Research, 69(8): 2882–2889.

Werbel, J. D., & Carter, S. M. 2002. The CEO's Influence on Corporate Foundation Giving. Journal of Business

Ethics, 40(1): 47–60.