The sensitivity of corporate social performance to corporate financial performance: A “time-based” agency theory perspective

Abstract

This study focuses on micro-level phenomena and time issues that have been traditionally neglected in both corporate governance and corporate social responsibility research. Drawing on agency theory concerning time-based managerial incentives (i.e. short term and long term), we investigate which managerial incentives for compensation drive the sensitivity of corporate social performance (CSP) to corporate financial performance (CFP). Using data for publicly listed Korean firms, we found a significant and positive relationship between CSP and CFP, with this relationship strengthened in firms with high managerial ownership but insignificant in those with high earnings-based compensation. Furthermore, we found that the interaction effects of CSP and high earnings-based compensation on CFP become positive in firms with high managerial ownership, indicating that the sensitivity between CSP and CFP is driven by long-term managerial incentives.

JEL Classification: M12, M14, G35

Keywords

Agency theory corporate financial performance corporate social performance corporate social responsibility managerial incentives

1. Introduction

Recent studies have found that corporate social responsibility (CSR) rating and ranking outcomes can significantly influence business analysts, investors, and rated firms (Chatterji and Toffel, 2010; Eccles et al., 2014), and market can response to corporate social performance and “signaling” for investors and shareholders (Ioannou and Serafeim, 2015). The accelerating rate of adoption of these practices has also provoked a debate about the nature of CSR and its long-term implications for business organizations. Over the last four decades, the debate of whether “it pays to be good” by examining the relationship between corporate social performance (CSP) and corporate financial performance (CFP) has endured and the outcomes of the related studies are mixed (Dixon-Fowler et al., 2013; Etzion, 2007; Guenther and Hoppe, 2014; Lee et al., 2016). Many management and strategy scholars have explored the relationship between CSP and CFP and provided abundant anecdotal evidence and successful business cases for CSR; however, the arguments on the relationship between CSP and CFP have yet to reach a conclusion (Guenther and Hoppe, 2014; Lee et al., 2015).

In this study, we use agency theory as our theoretical lens to examine the sensitivity between CSP and CFP. Agency theory argues that the strategic motivations and preferences of corporate managers are misaligned to those of shareholders. Agency problems can manifest through not-for-value-maximization or enhancement of investment choices and through managerial remuneration which is not tied to performance (Bebchuk and Fried, 2003; Ferrell et al., 2016; Shleifer and Vishny, 1989). As a result, managers may make decisions about investments and projects that, from the perspective of shareholders, are not the best. Traditionally, agency theory focuses on managers’ motivations and preferences to avoid risk (Gormley and Matsa, 2016) or large empire building (Jensen, 1986). However, Orlitzky et al. (2011) and McGuire et al. (2019) report that the lack of attention has been given the links between individual or micro-level phenomena (e.g. managerial compensation) and CSP.

When companies face CSR challenges from governments, consumers, investors, non-governmental organizations (NGOs), and other stakeholders, they may attempt to pursue long-term benefits by actively supporting and increasing CSP, resulting in positive CFP. The issue of whether companies face long-term versus short-term agency problems in the context of the sensitivity of CSP to CFP remains unexplored.

We aim to investigate whether the CSP-CFP sensitivity is valid in agency theory when it is combined with a “time issue” (hereafter, a “time-based” agency perspective). To examine whether CSP should be considered as part of an agency problem, we explore its underlying mechanism based on managerial incentives. Managerial incentives are a visible and important mechanism to direct managerial attention to specific goals and objectives having both financial and social performance implications (McGuire et al., 2019). Specifically, we investigate how short-term and long-term managerial incentive compensations affect the CSP-CFP sensitivity. The results will provide an opportunity to identify which managerial compensation incentives (short term and long term) drive CSP-CFP sensitivity and thus have implications for regulators and policy makers seeking CSP to mitigate agency problems between managers and external shareholders.

Most research states that high CSP results in positive impacts on CFP in the long-term perspectives because it is in line with the instrumental stakeholder hypothesis that CSP raises corporate reputation and mitigates regulatory pressure and activists or non-profit organizations’ concerns (Baron, 2001; Choi et al., 2010; Griffin and Mahon, 1997; Jones, 1995; Lev et al., 2010; Orlitzky et al., 2003; Roman et al., 1999; Waddock and Graves, 1997). In contrast, some studies have argued that high CSP results in negative impacts on CFP in the short-term perspectives because it intensifies the inefficient use of resources to meet the conflicting needs of various stakeholders (Aupperle et al., 1985; Brammer et al., 2006; Goldreyer and Diltz, 1999; Pava and Krausz, 1996). Thus, strategies for CSP which have a positive influence on CFP would be implemented with a long-term perspective (Peloza, 2006; Reid and Toffel, 2009).

From the perspective of time-based agency, there are competing hypotheses on the positive effect of CSP on CFP (Ortiz-de-Mandojana and Bansal, 2016; Wang and Bansal, 2012). First, high CSP may be a result of the effect of alignment on managers who are involved in CSR activities, wishing to align their interests with those of external shareholders, thus leading to increased long-term CFP (Anderson and Reeb, 2003; Cheng et al., 2012; Jensen and Meckling, 1976; Morck et al., 1988). On the other hand, high CSP may be a result of the effect of entrenchment on managers who engage in CSR activities to conceal opportunistic behaviors from stakeholders to improve short-term CFP (Carroll, 1979; Hemingway and Maclagan, 2004; Jensen and Meckling, 1976; McWilliams et al., 2006; Martínez-Ferrero et al., 2017; Prior et al., 2008). Although most previous studies find a positive relationship between CSP and CFP, they have not clarified whether or not this relationship is evidence of support for the instrumental stakeholder hypothesis. Previous studies have overlooked the fact that CSP-CFP sensitivity can be driven by managerial incentives based on time-based agency perspective. Thus, we use corporate shares owned by the management as the long-term managerial incentive, and earnings-based cash compensation as the short-term managerial incentive in order to add clarity to our hypotheses.

Many studies have identified that managerial incentives related to compensation are one of the causes of agency problems. With the prospect of long-term compensation, managers with low ownership may decrease corporate performance because they tend to pursue personal interests at the expense of external shareholders (McConnell and Servaes, 1990; Morck et al., 1988), whereas managers with high ownership may increase corporate performance because their interests can be aligned with those of external shareholders (Anderson and Reeb, 2003; Cheng et al., 2012; Jensen and Meckling, 1976; Morck et al., 1988). In the situation where highly paid managers of companies with earnings-based compensation schemes are paid compensation in the short term, they have a tendency to engage in opportunistic behavior to improve CFP in a bid to enhance their compensation (McWilliams et al., 2006; Martínez-Ferrero et al., 2017; Prior et al., 2008). Thus, it is necessary to explore which managerial incentives for compensation drive the CSP-CFP sensitivity and whether this relationship is in line with either the effect of “alignment” or the effect of “entrenchment” on managerial agency.

To investigate the effects of the time-based agency problem on the CSP-CFP sensitivity, this study examines (1) the relationship between CSP and CFP, (2) how this relationship is moderated by both high managerial ownership and earnings-based compensation, and (3) how the interaction between CSP and earnings-based compensation for CFP is moderated by high managerial ownership. In addition, to check the robustness of the results, we perform three tests. First, we analyze whether high CSP to achieve long-term goals is more dominant in family firms which are interested in long-term growth. Second, we also analyze the Korean large family business group Chaebol firm effect. As there are a relatively high number of large-sized Chaebol firms in the analysis, it is important to perform a robustness check to discern the effect on Chaebol-controlled firms. Third, we analyze whether managers with high ownership use high CSP to cover management’s opportunistic earnings. Managerial entrenchment occurs when managers use their discretion for their own interests, rather than for the interests of shareholders (Weisbach, 1988: 435).

This study found a significant and positive relationship between CSP and CFP. More importantly, this positive relationship is strengthened in firms with high managerial ownership but not in firms with high earnings-based compensation. The findings suggest that high CSP supports the instrumental stakeholder hypothesis, which accords with the effect of “alignment” on managers who use high CSP to align their interests with those of stakeholders, resulting in improved long-term CFP. Moreover, the effects of interaction on CSP and earnings-based compensation on CFP are switched, becoming positive in firms with high managerial ownership. These results show that CSP-CFP sensitivity is driven by long-term managerial incentives, implying that the long-term benefits of CSP to managers with high ownership outweigh the short-term benefits. Furthermore, the outcomes of the effects of family firms and Chaebol firms confirm that the instrumental stakeholder hypothesis is valid for family firms and non-Chaebol firms.

Consequently, this study provides important empirical evidence that the instrumental stakeholder hypothesis is valid in terms of the “time-based” agency perspective. As CSP is context-specific organizational actions and policies that take into account stakeholders’ expectations, this study provides important insights on the Korean business context to contribute to the debate on CSP-CFP sensitivity.

This study makes several important contributions to the literature. First, it introduces a “time” issue to examine CSP-CFP sensitivity using agency theory. Since a “time-based” agency problem taking managerial interests into account is a new and unexplored method through which to examine CSP-CFP sensitivity, this study provides new insights into the CSP-CFP debate by including long-term versus short-term managerial incentives. Second, this study further advances the CSR literature by suggesting that the instrumental stakeholder hypothesis can be further empirically validated by agency problems. Third, this study contributes to the management literature by considering the effects a family firm has on CSP-CFP sensitivity. Finally, this study provides further insight into CSP measures to enhance the validity of Korean CSP measures by providing three alternative measures (non-weighted, equally weighted, and stakeholder-weighted measures).

2. Background and hypotheses development

From the perspective of “time-based” agency, we explore why companies attempt to increase CSP, differentiating from the approach of previous CSP studies and focusing mainly on the CSP-CFP relationship. That is, we will demonstrate the internal organizational dynamics of how managers strategically respond to social pressure to manage the CSR challenges exerted by external stakeholders. Firms that rely very heavily on short-term investments tend to experience more volatile earnings. The legitimacy of the instrumental stakeholder hypothesis should be explained through a long-term agency perspective, because social pressures alleviated through CSR activities affect long-term CFP. In order to explain CSP-CFP sensitivity more clearly, it is important to understand managerial incentives to increase CSP.

2.1. Managerial incentives and CSP–CFP sensitivity

A number of studies argue that high CSP is positively associated with a firm’s investment activities, leading to higher CFP (known as the instrumental stakeholder hypothesis; Baron, 2001; Choi et al., 2010; Griffin and Mahon, 1997; Jones, 1995; Lev et al., 2010; Orlitzky et al., 2003; Roman et al., 1999; Waddock and Graves, 1997), whereas some studies suggest that high CSP only results in unnecessary extra costs, which in turn decrease CFP (Aupperle et al., 1985; Brammer et al., 2006; Goldreyer and Diltz, 1999; Pava and Krausz, 1996).

Baron (2001) argues that high CSP raises the competitive position of firms in an industry and ultimately increases corporate performance. In the same vein, Waddock and Graves (1997) and Orlitzky et al. (2003), using the Kinder Lydenberg Domoni (KLD) Index, find a positive association between CSP and CFP. Lev et al. (2010) show that charitable contributions are significantly associated with future revenue, suggesting that high CSP affects long-term financial performance. Using the Korea Economic Justice Institute (KEJI) Index, Choi et al. (2010) also found that CSP is positively associated with various CFP measures. On the other hand, Aupperle et al. (1985), Pava and Krausz (1996) and Goldreyer and Diltz (1999) argue that firms are reluctant to pay a premium for CSP because the costs for high CSP do not lead to an increase in CFP and investors are only interested in future high cash flow. Using UK data, Brammer et al. (2006) find a negative relationship between CSP and stock market returns, implying that companies’ excessive expenditure on CSP has a negative effect on CFP and that such companies are punished by the capital market. More recently, Chatterji and Toffel (2010) examined how firms might respond to CSP signals evidenced by CSR rating and ranking schemes. They found that firms change their performance in response to CSP signals. Using an event study methodology, Luffarelli and Awaysheh (2017) also found that shareholders value CSP using signals from the third parties to assess a firm’s CSP by confirming that the superior CSP is associated with higher firm performance.

Managers with high ownership, which can align their interests with those of external shareholders, would use high CSP to alleviate agency problems, resulting in an increase in long-term corporate performance (Baron, 2001; Choi et al., 2010; Lev et al., 2010; Orlitzky et al., 2003). However, corporate managers tend to undertake opportunistic behavior at the expense of external shareholders and thus might use high CSP to conceal opportunistic earnings in order to improve the firm’s short-term performance (McConnell and Servaes, 1990; Morck et al., 1988). Such entrenched managerial behavior regarding CSP may be more prevalent in firms with relatively high earnings-based compensation, which could serve as a catalyst, inciting management to inflate earnings for higher compensation (Gaver et al., 1995; Healy and Wahlen, 1999; Holthausen et al., 1995).

Taken together, if high CSP results from an effect of the alignment of managers’ interests, the CSP-CFP sensitivity will be strengthened in firms with high managerial ownership. However, if high CSP results from the effect of the entrenchment of managers, this sensitivity will be strengthened for firms with high earnings-based compensation. Thus, we propose the following hypotheses:

H1a: If the instrumental stakeholder hypothesis is in line with the effect of “alignment” on managerial agency, the sensitivity of CSP to CFP will be stronger for firms with high levels of managerial ownership.

H1b: If the instrumental stakeholder hypothesis is in line with the effect of “entrenchment” on managerial agency, the sensitivity of CSP to CFP will be stronger for firms with high earnings-based compensation.

2.2. Managerial ownership, earning-based compensation and CSP-CFP sensitivity

Recent research argues that managers use high CSP, which can lead to high short-term CFP, as a tool to conceal their opportunistic behavior (Hemingway and Maclagan, 2004; McWilliams et al., 2006; Martínez-Ferrero et al., 2017; Prior et al., 2008). Hemingway and Maclagan (2004) find that firms attempt to increase CSP to conceal illegal or unfair management practices. McWilliams et al. (2006) also find that managers use high CSP to achieve personal goals. In the context of the CSP-CFP sensitivity, Prior et al. (2008) found a positive relationship between earnings management and CSP among firms in regulatory industries such as gambling, tobacco, brewing, and arms manufacturers, suggesting that such firms exert pressure to increase CSP in order to conceal opportunistic behavior. More recently, Martínez-Ferrero et al. (2017) found that the positive relationship between CSP and CFP was strengthened in firms having a high degree of earnings management, suggesting that opportunistic earnings management can be covered by high CSP.

However, this stream of studies has overlooked the fact that CSP-CFP sensitivity can be driven by managerial “long-term” incentives. The legitimacy of the instrumental stakeholder hypothesis should be explained by the effect of alignment on managerial agency (Peloza, 2006; Reid and Toffel, 2009). Thus, if the long-term managerial benefits from high CSP are greater than the short-term benefits, the CSP-CFP sensitivity will be driven by long-term incentives for managers. Taken together, the interaction effect of CSP and earnings-based compensation on CFP will be positive for firms with high levels of managerial ownership. We propose the following hypothesis:

H2: If the sensitivity between CSP and CFP is driven by managerial long-term incentives, the effect of interaction on CFP of CSP and high earnings-based compensation will be positive for firms with high managerial ownership.

3. Sample selection and models

3.1. Sample selection

We obtain samples from stocks publicly traded on the Korea Composite Stock Price Index (KOSPI) to test our hypothesis. During the Asian financial crisis in the late 1990s, Korean firms underwent serious financial restructuring processes to improve transparency and accountability under the Korean government’s tough financial reform policy (Chang et al., 2017). Commonly, corporate social activities and engagement are also encouraged and practiced as a part of this process.

Panel A of Table 1 shows the selection process for the sample, and Panel B shows the distribution of the sample by year and by industry. In Panel A, we begin with 4560 companies that have been traded on the KOSPI during the period 2011-2016. We have selected the data period from 2011 to 2016 due to inconsistency in ratings of the KEJI Index which is used as a measure of CSP. The KEJI Index was the sum of the quantitative and qualitative evaluation scores of CSR activities until 2010. The quantitative evaluation score was out of 75, and quantitative evaluation was made on seven items: soundness, fairness, contribution to society, consumer protection, environmental protection, employee satisfaction, and contribution to economy. The score for qualitative evaluation is 25 points, which is the result of a survey on the company. However, in 2011, the KEJI Index calculation method was revised by deleting the score for qualitative evaluation and the item “contribution to economy” for quantitative evaluation. We first eliminate (1) 344 firms from the sample which have a closing month other than December, or which belong to the financial sector, to avoid any misleading results due to sample heterogeneity, and (2) 2259 firms not on the KEJI Index were removed from the sample. In order to increase the validity of the results, we eliminate (3) 28 firms whose financial data cannot be collected from the Korean FnGuide database (https://www.fnguide.com). We obtained the final sample of 1929 firms through these procedures. Panel B shows that 509 (26.4%) companies belonging to the chemical and chemical manufacturing industry have the highest frequency in the sample. The remaining samples were evenly distributed by industry.

Table 1.

Sample selection and distribution by industry and year.

Panel A: Sample selection
	Companies traded in the KOSPI market from 2011 to 2016							4560
(1)	Exclude companies that are not manufacturing industries and have closing months other than December							(344)
(2)	Exclude companies that do not have the KEJI Index disclosed by the Economic							(2259)
(3)	Justice InstituteExclude companies that cannot obtain financial data from FnGuide database							(28)
Final sample								1929
Panel B: Distribution by industry and year
Industry/year	2011	2012	2013	2014	2015	2016	Total	%
Food product and beverage	14	22	22	22	14	25	119	6.2
Pulp, paper, and paper products	7	10	9	12	6	7	51	2.6
Chemicals and chemical products	74	79	91	89	86	90	509	26.4
Other non-metallic mineral products	10	14	15	19	12	13	83	4.3
General construction	6	9	6	13	5	13	52	2.7
Basic metal products	29	28	30	32	30	34	183	9.5
Other machinery and equipment	6	24	16	21	13	17	97	5.0
Electrical and electronic equipment	11	31	34	33	28	29	166	8.6
Transportation equipment	6	31	24	20	19	22	122	6.3
Wholesales trade and commission trade	15	27	29	32	23	21	147	7.6
Furniture and other products	19	17	14	18	17	18	103	5.3
Professional services	43	48	46	22	21	48	228	11.8
Transportation	11	8	11	12	13	14	69	3.6
Total	251	348	347	345	287	351	1,929	100.0
%	13.0	18.0	18.0	17.9	14.9	18.2	100.0

KOSPI: Korea Composite Stock Price Index; KEJI: Korea Economic Justice Institute.

The industry classification of Panel B is based on Korea Standard Industry Classification (KSIC).

3.2. Models

3.2.1. CSP measures

Since 1991, the KEJI has published an index of listed firms with certain conditions, thereby identifying firms which are sound and well respected. The KEJI Index is composed of quantitative evaluations with the following six categories: soundness of capital structure (25), fairness of trade (20), contribution to communities (15), consumer protection and satisfaction (15), environmental protection (10), and employee satisfaction (15).

We use three alternative measures for CSP to compensate for some drawbacks the KEJI Index has for the purposes of this study. The first measure is the natural logarithm of the sum of non-weighted (NW) scores for six categories published by the KEJI. However, this measure has the possibility of distorting the importance of specific CSR activities because the weights of six subcategories are different each other. To address this drawback, consistent with Choi et al.’s (2010) approach, the second measure is the natural logarithm of the sum of equally weighted (EW) scores for six categories. After changing the different perfect score for individual categories to the same scores of 100, we calculate an equally weighted KEJI Index by the sum of the changed scores for six categories. For example, if a firm gains a score of 20 out of a maximum value of 25 in the soundness category, then a score of 20 becomes 80 (= 20 × (100/25)) because the maximum value of 25 is changed to 100. If a firm obtains a perfect score in all categories, the equally weighted KEJI Index is 600.¹ However, this measure is not considered an inherent risk across all industry–year observations. The importance of individual CSR categories may vary across the interests of stakeholders for every industry–year observation. To address this drawback, consistent with Akpinar et al.’s (2008) and Choi et al.’s (2010) approach, the last measure is the stakeholder-weighted (SW) KEJI Index. After classifying the KEJI Index of the sample by industry and year, we calculate the weight of each category by dividing the average score for individual categories for every industry–year observation by the average score for all categories for every industry–year observation. In sum, the stakeholder-weighted KEJI Index is measured by the sum of individual scores multiplied by the weight of each category for every industry–year observation.²

3.2.2. Model development

To test Hypotheses 1a, 1b, and 2, we develop equation (1) including control variables identified as determinants of CFP in previous studies (Aupperle et al., 1985; Cheng et al., 2012; Choi et al., 2010; Eccles et al., 2014; Orlitzky et al., 2003). Equation (1) employs return on assets (ROA), a measure of CFP as a dependent variable. The reason for using ROA is that earnings-based performance better reflects the efforts of management and plays a decisive role in determining management’s cash compensation figures.³ Cash compensation (salary or bonus) is usually based not on long-term performance (stock returns) but on short-term performance (reported earnings; Gul et al., 2003; Healy and Wahlen, 1999; Park, 2017).

As described above in CSP measures, three alternative measures for CSP (NW, EW, and SW) are used. If high CSP support the instrumental stakeholder hypothesis of increasing long-term performance by alleviating social pressures, the coefficients of CSP are all significant and positive (β₁ > 0). The MOWN variable employed as a long-term managerial incentive to increase CSP is a dummy variable, with a value of 1 if managerial ownership ranks among the top quartile of the sample, 0 otherwise.⁴ If more managers have more shares, the greater the interests of managers and external shareholders will be aligned, therefore managerial ownership (MOWN) will increase CFP (β₂ > 0). The COMP variable used as a short-term managerial incentive to increase CSP is also a dummy variable, with a value of 1 if the ratio of managerial cash compensation to total assets ranks among the top quartile of the sample and 0 otherwise. Because earnings-based compensation schemes serve as a catalyst for managers to increase CFP, high earnings-based management compensation (COMP) will increase CFP (β₃ > 0).

If high CSP results from the effect of alignment on managerial agency taking a long-term perspective, the CSP-CFP sensitivity will be strengthened in companies with high managerial ownership (β₄ > 0). However, if high CSP results from the effect of entrenchment on managerial agency taking a short-term perspective, the CSP-CFP sensitivity will be strengthened in firms with high earnings-based management compensation (β₅ > 0). Moreover, if the CSP-CFP sensitivity is driven by managerial long-term incentives, in line with the effect of alignment on managerial agency, the effect of the interaction between CSP and earnings-based compensation on CFP will be positive for firms with high levels of managerial ownership (β₆ > 0).

To eliminate the lagged effect of ROA, lagged ROA is included as Pre_ROA and this variable (Pre_ROA) will be positively related to ROA (β₇ > 0). To control firm size effects, the natural logarithm of total assets (SIZE) is included (β₈ > 0 or β₈ < 0). The higher the growth in sales and market value, the higher the financial performance of the company. Thus, Sales, Pre_Sales, and MB will have a positive relationship with ROA (β_{9, 10, 12} > 0). The larger the debt size, the higher the interest costs, the worse the profitability, and the higher the capital costs. Therefore, LEV will be negatively related to ROA (β₁₁ < 0). Research and development (RND) can contribute to an increase in CFP by increasing sales and creating added value, but excessive expenditure on research and development can be a cause of poor financial performance (β₁₃ > 0 or β₁₃ < 0)

\begin{array}{l} C F P (R O A) = α_{0} + β_{1} C S P (N W, E W, S W) + β_{2} M O W N + β_{3} C O M P \\ + β_{4} C S P (N W, E W, S W) \times M O W N + β_{5} C S P (N W, E W, S W) \times C O M P \\ + β_{6} C S P (N W, E W, S W) \times C O M P \times M O W N \\ + β_{7} P r e_R O A + β_{8} S I Z E + β_{9} S a l e s + β_{10} P r e_S a l e s + β_{11} L E V \\ + β_{12} M B + β_{13} R N D + ε \end{array}

(1)

where ROA is the earnings before interest and taxes divided by total assets, NW is the natural logarithm of the KEJI Index, EW is the equally weighted KEJI Index, SW is the stakeholder-weighted KEJI Index, MOWN takes the value of 1 if managerial ownership ranks among the top quartile of the sample and 0 otherwise, COMP takes the value of 1 if the ratio of management compensation to total assets ranks among the top quartile of the sample and 0 otherwise, Pre_ROA is the Lagged ROA, SIZE is the natural logarithm of total assets, Sales is the sales growth rate over the prior year, Pre_Sales is the Lagged Sales, LEV is the ratio of debt to total assets, MB is the ratio of market to book value, and RND is the R&D expenses deflated by total assets.

4. Empirical results

4. 1. Descriptive statistics

Table 2 shows the descriptive statistics for the variables of the sample. The mean (median) of ROA and Pre_ROA representing CFP are 4.3% (3.8%) and 4.1% (4.0%), respectively. The mean values of NW, EW, and SW for CSP are 4.126, 361.060, and 12.133, respectively. The average SIZE is 26.692 and the debt ratio is 38.5%. The averages of current (Sales) and previous sales growth rate (Pre_Sales) are 4.6% and 8.0%, respectively. Market value (MB) and research and development expenses (RND) account for 124.1% of book value and 0.6% of total assets, respectively.

Table 2.

Descriptive statistics.

Variable	Mean	Standard deviation	Minimum	Quartile			Maximum
Variable	Mean	Standard deviation	Minimum	25%	Median	75%	Maximum
ROA	0.043	0.047	–0.424	0.019	0.038	0.063	0.604
NW	4.126	0.049	3.944	4.094	4.127	4.160	4.279
EW	361.060	17.517	298.960	348.860	360.246	373.077	424.744
SW	12.133	0.634	9.966	11.724	12.166	12.568	14.408
Total compensation/assets	0.004	0.004	0.000	0.001	0.002	0.005	0.038
Management ownership	0.144	0.160	0.000	0.001	0.089	0.249	0.873
Pre_ROA	0.041	0.060	–0.725	0.017	0.040	0.067	0.567
SIZE	26.692	1.315	23.224	25.830	26.510	27.289	32.731
Sales	0.046	0.316	–0.731	–0.049	0.023	0.100	6.805
Pre_Sales	0.080	0.383	–0.998	–0.031	0.046	0.135	6.865
LEV	0.385	0.187	0.001	0.236	0.382	0.535	0.958
MB	1.241	1.126	0.161	0.608	0.932	1.448	13.284
RND	0.006	0.013	0.000	0.000	0.000	0.004	0.105

ROA: earnings before interest and taxes divided by total assets; NW: the natural logarithm of the KEJI Index; KEJI: Korea Economic Justice Institute; EW: the equally weighted KEJI Index; SW: the stakeholder-weighted KEJI Index; Management Ownership: the percentage of ownership owned by management; Pre_ROA: a lagged ROA; SIZE: the natural logarithm of total assets; Sales: the sales growth rate over the prior year; Pre_Sales: a lagged Sales; LEV: the ratio of debt to total assets; MB: the ratio of market to book value; RND: R&D expenses deflated by total assets.

Table 3 shows the results of the Pearson correlation among the variables. ROA has a significant positive correlation with NW, EW, and SW. These results indicate that the higher the CSP, the greater the CFP. This might be evidence supporting the instrumental stakeholder hypothesis, which social pressures alleviated by CSR activities lead to increased long-term performance. As expected, ROA shows a significant positive correlation with Pre_ROA, Sales, Pre_Sales, MB, and RND, while it shows a significant negative correlation with LEV.

Table 3.

Correlation matrix.

Variable	ROA	NW	EW	SW	Pre_ROA	SIZE	Sales	Pre_Sales	LEV	MB
ROA	1.00
NW	0.21***	1.00
EW	0.21***	0.97***	1.00
SW	0.18***	0.92***	0.81***	1.00
Pre_ROA	0.40***	0.18***	0.17***	0.17***	1.00
SIZE	0.02	0.15***	0.20***	0.08***	0.07***	1.00
Sales	0.08***	0.06**	0.06**	0.07***	–0.07***	–0.01	1.00
Pre_Sales	0.08***	0.07***	0.06**	0.07***	0.10***	0.02	0.20***	1.00
LEV	–0.27***	–0.09***	–0.06**	–0.11**	–0.31**	0.19**	0.05**	0.07***	1.00
MB	0.30***	0.16***	0.17***	0.11***	0.21***	0.09***	0.11***	0.10***	0.08***	1.00
RND	0.07***	0.26***	0.24***	0.25***	0.03	0.06**	0.06**	0.08***	–0.05**	0.15***

ROA: earnings before interest and taxes divided by total assets; NW: the natural logarithm of the KEJI Index; KEJI: Korea Economic Justice Institute; EW: the equally weighted KEJI Index; SW: the stakeholder-weighted KEJI Index; Pre_ROA: a lagged ROA; SIZE: the natural logarithm of total assets; Sales: the sales growth rate over the prior year; Pre_Sales: a lagged Sales; LEV: the ratio of debt to total assets; MB: the ratio of market to book value; RND: R&D expenses deflated by total assets.

***

and **Statistical significance at the 1% and 5% levels, respectively, using a two-tailed test.

4.2. Managerial ownership, compensation, and CSP-CFP sensitivity

Panels A–C in Table 4 show the results of equation (1) using NW, EW, and SW as CSP measures. To verify the sensitivity which occurs when excluding the largest industry clusters, Panel D shows the result of analyzing equation (1) after excluding firms in the chemicals and chemical products industry from the sample. Model (1) of Panel A shows that the coefficient of NW is significant and positive (β = 0.088, p < 0.001), and the coefficients of MOWN and NW × MOWN in model (2) are also both significant and positive (β = 0.393, 0.094, p < 0.05, 0.1). However, the coefficient of NW × COMP in model (2) is insignificant. These results show that positive CSP–CFP sensitivity is strengthened in companies with high managerial ownership, and thus, high CSP may support the instrumental stakeholder hypothesis from a long-term perspective. Moreover, the coefficient of NW × COMP × MOWN in model (3) is significant and positive (β = 0.003, p < 0.01). This result shows that the interaction effect of CSP and high earnings–based compensation on CFP becomes positive for firms with high managerial ownership, indicating that positive CSP-CFP sensitivity is driven by managerial long-term incentives, in line with the effect of alignment on managerial agency. It could be interpreted that the instrumental stakeholder hypothesis is valid from the perspective of “time-based” agency. The results of Panel A are similar to those of Panels B and C, indicating that they are robust to the equal weight on the sub-items of the KEJI Index and the weights reflecting stakeholders’ interests on them. The results in Panel D are the same as those in Panels A to C. In conclusion, positive CSP-CFP sensitivity can be evidence of the instrumental stakeholder hypothesis. These results support hypotheses H1a and H2.

Table 4.

Managerial ownership, compensation, and the CSP-CFP sensitivity.

Panel A CSP measure = NW ROA = α₀ + β₁NW + β₂MOWN + β₃COMP + β₄NW × MOWN + β₅NW × COMP + β₆NW × COMP × MOWN + β₇Pre_ROA + β₈SIZE + β₉Sales + β₁₀Pre_{_}Sales + β₁₁LEV + β₁₂MB + β₁₃RND + ε
Variable	Model 1		Model 2		Model 3
NW	0.088	(3.988)***	0.062	(2.533)**	0.063	(2.565)***
MOWN			0.393	(1.983)**	0.388	(1.962)*
COMP			–0.127	(–0.623)	–0.095	(–0.466)
NW × MOWN			0.094	(1.963)*	0.092	(1.920)*
NW × COMP			0.031	(0.616)	0.022	(0.438)
NW × COMP × MOWN					0.003	(2.582)***
Pre_ROA	0.205	(11.680)***	0.204	(11.644)***	0.203	(11.582)***
SIZE	0.001	(0.635)	0.001	(0.144)	–0.001	(–0.163)
Sales	0.011	(3.535)***	0.011	(3.452)***	0.011	(3.426)***
Pre_Sales	0.003	(0.920)	0.003	(0.971)	0.003	(0.934)
LEV	–0.054	(–9.452)***	–0.056	(–9.700)***	–0.057	(–9.854)***
MB	0.011	(11.419)***	0.011	(11.354)***	0.011	(11.357)***
RND	–0.067	(–0.900)	–0.070	(–0.931)	–0.082	(–1.097)
Intercept	–0.318	(–3.678)***	–0.212	(–2.124)**	–0.207	(–2.074)**
Industry dummies	Yes		Yes		Yes
Year dummies	Yes		Yes		Yes
Adjusted R²	26.1%		26.3%		26.5%
F value	28.301***		24.746***		24.215***
Panel B CSP measure = EW ROA = α₀ + β₁EW + β₂MOWN + β₃COMP + β₄EW × MOWN + β₅EW × COMP + β₆EW × COMP × MOWN + Controls + ε
Variable	Model 1		Model 2		Model 3
EW	0.096	(4.358)***	0.076	(2.992)***	0.076	(2.996)***
EW × MOWN			0.610	(1.358)	0.585	(1.304)
EW × COMP			0.329	(0.705)	0.227	(0.485)
EW × COMP × MOWN					0.075	(2.253)***
Controls	Yes		Yes		Yes
Adjusted R²	26.3%		26.4%		26.6%
F value	28.469***		24.790***		24.244***
Panel C CSP measure = SW ROA = α₀ + β₁SW + β₂MOWN + β₃COMP + β₄SW × MOWN + β₅SW × COMP + β₆SW × COMP × MOWN + Controls + ε
Variable	Model 1		Model 2		Model 3
SW	0.062	(2.748)***	0.039	(1.614)*	0.042	(1.617)*
SW × MOWN			1.027	(2.472)**	0.967	(2.329)**
SW × COMP			0.072	(0.172)	–0.032	(–0.076)
SW × COMP × MOWN					0.080	(2.672)***
Controls	Yes		Yes		Yes
Adjusted R²	25.8%		26.0%		26.3%
F value	27.847***		24.419***		23.919***
Panel D Sensitivity test results by excluding the largest industry clustersROA = α₀ + β₁CSP + β₂MOWN + β₃COMP + β₄CSP × MOWN + β₅CSP × COMP + β₆CSP × COMP × MOWN + Controls + ε
Variable	CSP measure = NW		CSP measure = EW		CSP measure = SW
NW	0.064	(2.370)**
NW × MOWN	0.059	(1.630)*
NW × COMP × MOWN	0.004	(3.048)***
EW			0.081	(2.746)***
EW × MOWN			0.386	(0.682)
EW × COMP × MOWN			0.111	(3.002)***
SW					0.046	(1.639)***
SW × MOWN					0.852	(1.616)*
SW × COMP × MOWN					0.114	(3.083)***
Controls	Yes		Yes		Yes
Adjusted R²	22.5%		22.5%		22.3%
F value	15.612***		15.094***		14.940***

CSP: corporate social performance; CFP: corporate financial performance; NW: the natural logarithm of the KEJI Index; KEJI: Korea Economic Justice Institute; ROA: earnings before interest and taxes divided by total assets; MOWN: managerial ownership; COMP: compensation; Pre_ROA: a lagged ROA; SIZE: the natural logarithm of total assets; Sales: the sales growth rate over the prior year; Pre_Sales: a lagged Sales; LEV: the ratio of debt to total assets; MB: the ratio of market to book value; RND: R&D expenses deflated by total assets; EW: the equally weighted KEJI Index; SW: the stakeholder-weighted KEJI Index.

Panels A–C show the results of equation (1) containing NW, EW, and SW as CSP measures, respectively. To verify the sensitivity of excluding the largest industry clusters, Panel D shows the result of equation (1) for the sample from which the firms in chemicals and chemical products industry are excluded. These findings show that the results of Panels A–C were not driven by the largest industry clusters. The reports of the results for MOWN, COMP, and CSP measures × COMP are omitted for brevity. Standard errors are in parentheses.

p < 0.1, **p < 0.05, and ***p < 0.01 levels, two-tailed coefficient test (N = 1929).

4.3. Endogeneity of CSP and CFP

We tested two-stage least squares (2SLS) models to address an issue of potential endogeneity among our variables and moderators. Firms with good financial performance can afford to participate in CSR activities, which can lead to higher CSP (Ullman, 1985; Waddock and Graves, 1997), while firms with higher CSP can increase their reputation, lower regulatory pressures, and sooth concerns from activists and NGOs, eventually leading to higher CFP (Baron, 2001; Lev et al., 2010; Orlitzky et al., 2003). Table 5 shows the results of analyses of 2SLS regressions that are based on Choi et al.’s (2010) approach. In the first step, we analyze equation (2) that includes CSP as a dependent variable, ROA as the endogenous variable, and the prior year’s CSP (Pre_CSP) and firm size (SIZE) as instrumental variables.⁵ To gain the lagged variables, we lost 618 observations from the original sample, which results in 1311 observations for 2SLS regressions. In the second step, we analyze the following equation (3), which CSP in equation (1) is replaced by the predicted CSP (CSPHAT: NWHAT, EWHAT, SWHAT) estimated in equation (2)

C S P = α_{0} + β_{1} R O A + β_{2} P r e_C S P + β_{3} S I Z E + ε

(2)

\begin{array}{l} R O A = α_{0} + β_{1} C S P H A T + β_{2} M O W N + β_{3} C O M P + β_{4} C S P H A T \times M O W N \\ + β_{5} C S P H A T \times C O M P + β_{6} C S P H A T \times C O M P \times M O W N \\ + C o n t r o l s + ε \end{array}

(3)

Table 5.

Results of 2SLS regression.

Panel A: First stageCSP = α₀ + β₁ROA + β₂Pre_CSP + β₃SIZE + ε
Variable	(1)CSP = Pre_NW		(2)CSP = Pre_EW		(3)CSP = Pre_SW
ROA	0.101	(4.387)***	0.117	(4.984)***	0.068	(3.108)***
Pre_NW	0.887	(17.787)***
Pre_EW			0.362	(14.345)***
Pre_SW					1.047	(20.678)***
SIZE	0.077	(3.302)***	0.130	(5.363)***	0.035	(1.692)*
Adjusted R²	35.0%		31.0%		41.5%
F value	36.218***		30.448***		47.373***
Panel B: Second stageROA = α₀ + β₁CSPHAT + β₂MOWN + β₃COMP + β₄CSPHAT × MOWN + β₅CSPHAT × COMP + β₆CSPHAT × COMP × MOWN + Controls + ε
Variable	(1)CSP = NWHAT		(2)CSP = EWHAT		(3)CSP = SWHAT
NWHAT	0.213	(7.010)***
NWHAT × MOWN	0.050	(1.823)*
NWHAT × COMP	–0.003	(–0.130)
NWHAT × COMP × MOWN	0.049	(1.658)*
EWHAT			0.305	(8.575)***
EWHAT × MOWN			1.828	(2.273)**
EWHAT × COMP			0.705	(0.830)
EWHAT × COMP × MOWN			0.043	(1.671)*
SWHAT					0.129	(3.842)***
SWHAT × MOWN					0.530	(0.816)
SWHAT × COMP					0.783	(1.238)
SWHAT × COMP × MOWN					0.055	(1.829)*
Controls	Yes		Yes		Yes
Adjusted R²	50.0%		51.8%		48.9%
F value	45.723***		45.868***		40.747***
N	1310		1310		1310

2SLS: two-stage least squares; CSP: corporate social performance; ROA: earnings before interest and taxes divided by total assets; SIZE: the natural logarithm of total assets; MOWN: managerial ownership; COMP: compensation; NW: the natural logarithm of the KEJI Index; KEJI: Korea Economic Justice Institute; EW: the equally weighted KEJI Index; SW: the stakeholder-weighted KEJI Index.

Panel A shows the results of equation (2) that includes CSP as a dependent variable, ROA as the endogenous variable, and Pre_CSP as instrumental variables. Panel B shows those of equation (3), which CSP in equation (1) is replaced by the predicted CSP (CSPHAT: NWHAT, EWHAT, SWHAT) estimated in Panel A. Standard errors are in parentheses.

p < 0.1, **p < 0.05, and ***p < 0.01 level, two-tailed coefficient test.

Panel A presents the results of analyses for equation (2) and Panel B shows those for equation (3). In Panel A, the coefficients on ROA and lagged CSP (Pre_NW, Pre_EW, Pre_SW) are all significant and positive (β = 0.101, 0.117, 0.068, 0.887, 0.362, 1.047, all p < 0.01). In Panel B, the coefficients on CSPHAT (NWHAT, EWHAT, SWHAT) as well as CSPHAT (NWHAT, EWHAT, SWHAT) × COMP × MOWN are all significant and positive (β = 0.213, 0.305, 0.129, 0.049, 0.043, 0.055, all p < 0.1). These results show that the results of Table 4 are robust to the consideration of endogeneity between CSP and CFP.

4.4. Further analyses

4.4. 1. Family firm effects

In the majority of countries, family-run firms account for more than 50% of all firms (Gomez-Mejia et al., 2007, 2010, 2011; La Porta et al., 1999). Unlike firms dominated by large numbers of shareholders, family firms are companies in which family assets are invested. Notably, family members are long-term investors throughout several generations, and they have been characterized as taking their family reputation seriously (Anderson and Reeb, 2003). Previous studies argue that family firms have a positive effect on firm value because the family ownership structure, where ownership and control are not separated, alleviates agency problems (Anderson and Reeb, 2003; Martínez et al., 2007). Thus, it is necessary to analyze how CSP is used in Korean family firms. Although a definition of family firms is crucial in the development of empirical models, there has been no united definition, due to the differences in the institutions or in the legislation throughout various countries. However, many previous studies have suggested three alternative definitions for family firms as follows.⁶ The first definition is that the family firm is a firm with two or more family–member managers (Anderson and Reeb, 2003; McConaughy et al., 1998). The second definition is that it is a firm where some managers are family members in the second or more generation from the founder (Villalonga and Amit, 2006; Westhead and Cowling, 1998; Westhead et al., 2002). The third definition is that it is a firm in which one of the family members is the CEO (Martínez et al., 2007; Maury, 2006; Smith and Amoako Abu, 1999).

Table 6 shows how the results of Table 4 depend on the types of family firms. Panels A to C show the results of analyses including NW, EW, and SW as CSP measures, respectively. In Panel A, the coefficients of NW × COMP × MOWN are significantly positive only for firms with two or more family–member managers, family–member managers in the second or more generation, and a family–member CEO (β = 0.004, 0.003, 0.003, all p < 0.05). We also consider to test the equality of the coefficients between these two groups for claiming that family firm matters (Wooldridge, 2016). To address this issue, we conduct the Chow-test for all models (Chow, 1960). It shows whether the coefficient of three-way interactive term is equal across groups. Under each panel, we provide the results that the difference between coefficients of two regressions is significant. These results suggest that family–member managers use CSP for long-term purposes, and thus, the instrumental stakeholder hypothesis is valid for family firms. Panels B and C show the same results as Panel A.

Table 6.

Family firm effects.

Panel A CSP measure = NW ROA = α₀ + β₁NW + β₂MOWN + β₃COMP + β₄NW × MOWN + β₅NW × COMP + β₆NW × COMP × MOWN + Controls + ε
Variable	Two or more family members				Succession management				Family member CEO
Variable	More than one		Less than one		Over second generation		Under second generation		Family		Non-family
NW	0.097	(2.044)**	0.033	(1.265)	0.070	(2.309)**	0.066	(1.710)*	0.068	(2.066)**	0.058	(1.723)*
NW × MOWN	0.031	(1.614)*	0.123	(2.015)**	0.062	(1.136)	0.140	(1.574)	0.076	(1.699)*	0.064	(0.467)
NW × COMP × MOWN	0.004	(2.232)**	0.001	(0.777)	0.003	(2.207)**	0.001	(0.616)	0.003	(2.106)**	0.004	(1.349)
Controls	Yes		Yes		Yes		Yes		Yes		Yes
Adjusted R²	18.4%		38.5%		28.7%		27.0%		20.8%		43.1%
F value	6.967***		25.437***		15.912***		11.398***		13.118***		15.089***
H₀: β₆^Family = β₆^Non-Family	5.173**				4.824**				6.975***
N	798		1,131		1,112		817		1,389		540
Panel B CSP measure = EW ROA = α₀ + β₁EW + β₂MOWN + β₃COMP + β₄EW × MOWN + β₅EW × COMP + β₆EW × COMP × MOWN + Controls + ε
Variable	Two or more family members				Succession management				Family member CEO
Variable	More than one		Less than one		Over second generation		Under second generation		Family		Non-family
EW	0.109	(2.482)**	0.058	(1.923)*	0.100	(2.990)***	0.067	(1.724)*	0.083	(2.544)**	0.076	(1.904)*
EW × MOWN	0.007	(0.010)	0.840	(1.567)	0.256	(0.441)	0.657	(0.904)	0.397	(0.733)	0.232	(0.273)
EW × COMP × MOWN	0.112	(2.205)**	0.023	(0.635)	0.081	(2.137)**	0.022	(0.448)	0.076	(2.044)**	0.054	(1.115)
Controls	Yes		Yes		Yes		Yes		Yes		Yes
Adjusted R²	18.8%		38.4%		28.6%		27.0%		20.7%		43.1%
F value	7.369***		25.303***		16.315***		11.433***		13.516***		15.073***
H₀: β₆^Family = β₆^Non-family	5.516**				5.131**				7.376***
N	798		1131		1112		817		1389		540
Panel C CSP measure = SW ROA = α₀ + β₁SW + β₂MOWN + β₃COMP + β₄SW × MOWN + β₅SW × COMP + β₆SW × COMP × MOWN + Controls + ε
Variable	Two or more family members				Succession management				Family member CEO
Variable	More than one		Less than one		Over second generation		Under second generation		Family		Non-family
SW	0.004	(1.673)*	0.002	(0.896)	0.003	(1.670)*	0.003	(1.062)	0.003	(1.691)*	0.004	(1.460)
SW × MOWN	0.005	(0.774)	0.012	(2.487)**	0.008	(1.775)*	0.011	(1.526)	0.008	(1.985)**	0.002	(0.206)
SW × COMP × MOWN	0.001	(2.249)**	0.001	(0.936)	0.001	(2.193)**	0.001	(0.652)	0.001	(2.107)**	0.001	(1.306)
Controls	Yes		Yes		Yes		Yes		Yes		Yes
Adjusted R²	17.5%		38.4%		28.1%		26.8%		20.4%		42.8%
F value	6.829***		25.337***		15.997***		11.278***		13.299***		14.922***
H₀: β₆^Family = β₆^Non-Family	5.752**				5.357**				7.649***
N	798		1131		1112		817		1389		540

CSP: corporate social performance; NW is the natural logarithm of the KEJI Index; KEJI: Korea Economic Justice Institute; ROA: earnings before interest and taxes divided by total assets; MOWN: managerial ownership; COMP: compensation; CEO: chief executive officer; EW: the equally weighted KEJI Index; SW: the stakeholder-weighted KEJI Index.

Panels A–C show the results of equation (1) containing NW, EW, and SW as CSP measures, respectively. Family firms were defined as three types, and the sample was divided into groups belonging to family firms and those not. The reports of the results for MOWN, COMP, and CSP measures × COMP are omitted for brevity. Standard errors are in parentheses.

p < 0.1, **p < 0.05, and ***p < 0.01 levels, two-tailed coefficient test.

4.4.2. Chaebol firm effects

A large Korean business group, also known as Chaebol, is described as “a group of companies of which more than 30% of shares are owned by the group’s controlling shareholder and its affiliated companies” (Korea Fair Trade Commission (KFTC), 2015). In the Korean business context, Chaebols indicate large family business groups such as Samsung, Hyundai, and LG groups. Korean Chaebol firms under a shareholding agreement that allows the parent company to control the affiliates have published CSR reports to demonstrate their CSR engagement activities and CSP. However, they face the principal –principal conflicts of a controlling shareholder and minority shareholders and tend to engage in “tunneling” of converting the wealth of external stakeholders to their wealth (Baek et al., 2006; Park, 2017). Thus, they are likely to be active in CSR activities in order to conceal their opportunistic behavior in the short term. To check the sensitivity between Chaebol firms and non-Chaebol firms, we split the sample firms into Chaebol and non-Chaebol firms for further analysis. Table 7 shows how the results of Table 4 depend on Chaebol firms and non-Chaebol firms. All coefficients of three-way interactive terms are significantly positive only for Chaebol firms. These results suggest that Chaebol firms use CSP for short-term purposes, and thus, the instrumental stakeholder hypothesis is valid for non-Chaebol firms.

Table 7.

Chaebol firm effects.

Variable	NW				EW				SW
	Chaebol		Non-Chaebol		Chaebol		Non-Chaebol		Chaebol		Non-Chaebol
NW	0.055	(1.808)*	0.075	(2.244)**
NW × MOWN	–0.033	(–0.337)	0.108	(1.919)*
NW × COMP × MOWN	0.002	(0.132)	0.012	(2.310)**
EW					0.001	(1.723)*	0.001	(2.854)***
EW × MOWN					0.001	(0.428)	0.001	(1.293)
EW × COMP × MOWN					0.001	(0.176)	0.001	(2.233)**
SW									0.003	(1.057)	0.004	(1.823)*
SW × MOWN									–0.002	(–0.303)	0.010	(2.352)**
SW × COMP × MOWN									0.001	(0.077)	0.001	(2.398)**
Controls	Yes		Yes		Yes		Yes		Yes		Yes
Adjusted R²	52.0%		24.6%		52.1%		24.6%		51.9%		24.2%
F value	15.858***		15.663***		15.888***		15.727***		15.764***		15.408***
H₀: β₆^Chaebol = β₆^Non-Chaebol	4.460**				4.927**				5.146**
N	454		1,475		454		1,475		454		1,475

NW: the natural logarithm of the KEJI Index; KEJI: Korea Economic Justice Institute; EW: the equally weighted KEJI Index; SW: is the stakeholder-weighted KEJI Index; MOWN: managerial ownership; COMP: compensation; CEO: chief executive officer.

The sample was divided into groups belonging to Chaebol firms and those not. Standard errors are in parentheses.

p < 0.1, **p < 0.05, and ***p < 0.01 level, two-tailed coefficient test.

4.4.3. Earnings management, managerial ownership, and the CSP-CFP sensitivity

Managers with a high degree of ownership may use high CSP to conceal opportunistic behavior in order to increase CFP. These managers are likely to use the influence stemming from ownership to inflate earnings opportunistically (EM). Thus, we set equation (4) to analyze how the effect of earnings management on the CSP–CFP sensitivity is moderated by managerial ownership

\begin{array}{l} R O A = α_{0} + β_{1} C S P + β_{2} E M + β_{3} M O W N + β_{4} C S P \times E M + β_{5} C S P \times M O W N \\ + β_{6} C S P \times E M \times M O W N + C o n t r o l s + ε \end{array}

(4)

We use discretionary accruals estimated from the modified Jones model (Dechow et al., 1995)⁷ to measure the level of earnings management (EM). EM is a dummy variable which takes a value of 1 when discretionary accruals have a positive value, otherwise 0. As discretionary accruals increase, earnings are inflated, so positive discretionary accruals (EM) will be positively related to CFP (β₂ > 0). If CSP results from the effect of “alignment” on managers, positive CSP-CFP sensitivity will be weakened in firms with positive discretionary accruals (β₄ < 0). However, if CSP is a result of the effect of “entrenchment” on managers, positive CSP-CFP sensitivity will be strengthened in firms with positive discretionary accruals (β₄ > 0). If managers with high ownership use CSP as a tool to conceal their opportunistic earnings, the positive interaction of CSP with EM on CFP will be strengthened in firms with high managerial ownership (β₆ > 0). However, if managers with high ownership use CSP as a tool to align their interests with those of external stakeholders, the positive interaction of CSP with EM on CFP will be weakened in firms with high managerial ownership (β₆ < 0).

Table 8 shows that the coefficients of EM in Models 1 and 3 are significant and positive (β = 0.258, 0.805, p < 0.1 or 0.05), indicating that CFP increases with managerial EM. However, the coefficients of NW × EM and SW × EM are both significant and negative (β = −0.061, −0.706, both p < 0.1), suggesting that CSP is a result of the effect of alignment on managers. In addition, to investigate whether managers with high ownership use high CSP to achieve their opportunistic goals, we re-estimate our models including interactions between three variables (CSP, EM, MOWN; see Table 8). The coefficients on the three-way interaction terms (CSP × EM × MOWN) are also all negative but insignificant, suggesting that managerial influence stemming from high ownership is not used to achieve opportunistic goals, and the instrumental stakeholder hypothesis is valid again in terms of agency theory.

Table 8.

Managerial ownership, earnings management and the CSP–CFP sensitivity.

ROA = α₀ + β₁CSP + β₂EM + β₃MOWN + β₄CSP × EM + β₅CSP × MOWN + β₆CSP × EM × MOWN + Controls + ε
Variable	CSP measure = NW		CSP measure = EW		CSP measure = SW
NW	0.093	(3.154)***
MOWN	0.410	(2.126)**
EM	0.258	(1.667)*
NW × MOWN	0.099	(2.110)**
NW × EM	–0.061	(–1.605)*
NW × EM × MOWN	–0.001	(–0.364)
EW			0.101	(3.269)***
MOWN			0.696	(1.606)
EM			0.577	(1.429)
EW × MOWN			0.671	(1.545)
EW × EM			–0.477	(–1.179)
EW × EM × MOWN
SW					0.074	(2.380)**
MOWN					0.115	(2.608)**
EM					0.805	(2.144)**
SW × MOWN					1.035	(2.538)**
SW × EM					–0.706	(–1.874)*
SW × EM × MOWN					–0.007	(–0.261)
Controls	Yes		Yes		Yes
Adjusted R²	27.3%		27.3%		27.1%
F value	25.145***		25.116***		24.897***

CSP: corporate social performance; CFP: corporate financial performance; ROA: earnings before interest and taxes divided by total assets; EM: earnings management; MOWN: managerial ownership; NW: the natural logarithm of the KEJI Index; KEJI: Korea Economic Justice Institute; EW: the equally weighted KEJI Index; SW: the stakeholder-weighted KEJI Index.

EM is a dummy variable which takes a value of 1 when discretionary accruals estimated from the modified Jones model (Dechow et al., 1995) are a positive value, otherwise 0. The reports of the results for CSP measures × COMP are omitted for brevity. Standard errors are in parentheses.

p < 0.1, **p < 0.05, and ***p < 0.01 level, two-tailed coefficient test.

5. Conclusion

In agency theory, arguably managers have competing effects on positive CSP-CFP sensitivity, and these effects have been identified in most studies. While the effect of alignment on managers who use CSP to align managerial interests with external shareholders exists (Anderson and Reeb, 2003; Cheng et al., 2012; Jensen and Meckling, 1976; Morck et al., 1988), there is also the effect of entrenchment on managers who invest to increase CSP to conceal opportunistic behavior for their own interests (Carroll, 1979; Hemingway and Maclagan, 2004; Jensen and Meckling, 1976; McWilliams et al., 2006; Martínez-Ferrero et al., 2017; Prior et al., 2008).

The purpose of this study is to shed some light on the validity of the instrumental stakeholder hypothesis from a time-based agency perspective in examining the CSP-CFP sensitivity. Our study shows that there is positive CSP-CFP sensitivity, which is strengthened in companies with high managerial ownership but which is insignificant in those with high earnings-based compensation. These results suggest that the instrumental stakeholder hypothesis is further validated from the perspective of long-term agency. In addition, the positive interaction effects of CSP and high earnings-based compensation on CFP are strengthened in companies with high managerial ownership. These results indicate that the long-term benefits of high CSP outweigh the short-term benefits, and the results can be evidence of the legitimacy of the instrumental stakeholder hypothesis that social pressures alleviated by CSR activities lead to an increase in corporate long-term performance. Furthermore, the above results are more evident in family firms and clarify the fact that the instrumental stakeholder hypothesis is also valid for family firms. These results are the same in the analyses that control endogeneity between CSP and CFP.

This study contributes to the literature in several ways. First, our study sheds light on the sensitivity between CSP and CFP from an agency theory perspective. Traditionally, agency theory assumes opportunistic behavior of managers without considering managerial short- and long-term interests. As our response to Orlitzky et al. (2011) and McGuire et al.’s (2019) call for study on the links between individual or micro-level managerial phenomena and CSR, we extend the agency theory by incorporating time-based (short- or long-term) managerial incentives on CSP and CFP. More specifically, we examine the effect of time-based managerial incentives on CSP and financial performance, based on agency theory. The time factor (short term and long term) plays an important role in managerial decision making for CSR activities, hence understanding how short-term and long-term managerial incentives impact the sensitivity between CSP and CFP is at the core of validating the instrumental stakeholder hypothesis. Second, we extend existing literature by investigating the relevance of agency theory to CSP and its implications for the relationship between managerial compensation and CSP. More specifically, managers whose remuneration is heavily weighted toward short-term incentives may be less inclined to invest in high CSP, while those whose remuneration is more heavily weighted toward long-term incentives may have greater motivation to invest in high CSP. By providing empirical evidence, we provide a more nuanced understanding of how and why managerial compensation affects CSP. Third, as CSP is “context-specific organizational actions and policies that take into account stakeholders’ expectations and the triple bottom line of economic, social, and environmental performance” (Aguinis and Glavas, 2012: 855), this study enriches our understanding on managerial compensation and its design of decision makers relating to CSR engagement and investment. Our findings indicate that it would be beneficial for companies to develop compensation plans and schemes that can align the interests of managers and shareholders in order for CSR management activities to become catalysts in mitigating agency problems.

The findings from this study have important managerial implications. Managers face more pressure to demonstrate short-term returns. Many managers have difficulty investing in major projects that accrue benefits in the long term, while some managers invest in long-term activities that benefit stakeholders and society (Baron, 2001; Choi et al., 2010; Griffin and Mahon, 1997; Jones, 1995; Lev et al., 2010; Orlitzky et al., 2003). There is often a contest between what is good for the firm in the short term and what is good for stakeholders and society in the long term. Understanding the decisions requires that attention be paid to the managers’ incentives mechanisms, more specifically compensation plans. Some companies such as Alcoa, Intel, and Xcel Energy have adopted the integration of CSR criteria in executive compensation, what they call “CSR contracting” (Flammer et al., 2019). Our results can help firms to address whether inclusion of CSR criteria in compensation plans or managerial incentives promotes general CSR issues and engagement. As our study finds out, high managerial ownership plays an important role in pursuing CSR activities, resulting in superior financial returns. In particular, long-term managerial incentives encourage executives and decision makers to gain long-term benefits of CSP. We notice that most incentives and compensation plans do not explicitly state or target corporate social issues. It may be useful for providing clear objectives or guidance to firms when designing the compensation plan and structure for executives or managers if CSR has an important role in the company (Brown-Liburd and Zamora, 2015).

Although the study makes important contributions, it also has several limitations. As our sample and CSP measures using the KEJI Index are based in the context of Korean businesses, our study is limited to generalizing our findings without testing further in other countries. However, our findings are based on an in-depth single-country study, and what we provide is a deeper understanding of CSP-CFP linkages in a specific institutional environment. Therefore, future studies may benefit from examining whether the instrumental stakeholder hypothesis can be found in other institutional contexts such as North America, Europe, and in other Asian countries. Finally, our sample is limited to a relatively short time period, thus for high research validity, future studies need to examine whether our hypotheses are confirmed over a longer time period.

Footnotes

Final transcript accepted 11 March 2020 by Sue Wright (AE Accounting).

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study received partial financial support from Soonchunhyang University.

ORCID iD

Ki-Hoon Lee

Notes

References

Aguinis

Glavas

(2012) What we know and don’t know about corporate social responsibility: A review and research agenda. Journal of Management 38: 932–968.

Akpinar

Jiang

Gomez-Mejia

(2008) Strategic use of CSR as a signal for good management. IE Business School Working Paper, WP08-25. Available at: http://latienda.ie.edu/working_papers_economia/WP08-25.pdf

Anderson

Reeb

(2003) Founding family ownership and firm performance: Evidence from the S&P 500. The Journal of Finance 58: 1301–1328.

Aupperle

Carroll

Hatfield

(1985) An empirical examination of the relationship between corporate social responsibility and profitability. The Academy of Management Journal 28: 446–463.

Baek

J-S

Kang

J-K

Lee

I-M

(2006) Business groups and tunneling: Evidence from private securities offerings by Korean Chaebols. The Journal of Finance 61: 2415–2449.

Baron

(2001) Private politics, corporate social responsibility and integrated strategy. Journal of Economics and Management Strategy 10: 7–45.

Bebchuk

Fried

(2003) Executive compensation as an agency problem. Journal of Economic Perspectives 17: 71–92.

Brammer

Brooks

Pavelin

(2006) Corporate social performance and stock returns: UK evidence from disaggregate measures. Financial Management 35: 97–116.

Brown-Liburd

Zamora

(2015) The role of corporate social responsibility (CSR) assurance in investors’ judgments when managerial pay is explicitly tied to CSR performance. Auditing: A Journal of Practice & Theory 34: 75–96.

10.

Carroll

(1979) A three-dimensional conceptual model of corporate performance. Academy of Management Review 4: 497–505.

11.

Chang

Park

, et al. (2017) Exploring the relationship between board characteristics and CSR: Empirical evidence from Korea. Journal of Business Ethics 140: 225–242.

12.

Chatterji

Toffel

(2010) How firms respond to being rated. Strategic Management Journal 31: 917–945.

13.

Cheng

Zhu

(2012) Managerial ownership, board monitoring and firm performance in a family-concentrated corporate environment. Accounting & Finance 52: 1061–1081.

14.

Choi

Kwark

Choe

(2010) Corporate social responsibility and corporate financial performance: Evidence from Korea. Australian Journal of Management 35: 291–311.

15.

Chow

(1960) Tests of equality between sets of coefficients in two linear regressions. Econometrica: Journal of the Econometric Society 28: 591–605.

16.

Dechow

Sloan

Sweeney

(1995) Detecting earnings management. The Accounting Review 70: 193–225.

17.

Dixon-Fowler

Slater

Johnson

(2013) Beyond “does it pay to be green?” A meta-analysis of moderators of the CEP-CFP relationship. Journal of Business Ethics 112: 353–366.

18.

Eccles

Ioannou

Serafeim

(2014) The impact of corporate sustainability on organizational processes and performance. Management Science 60: 2835–2857.

19.

Etzion

(2007) Research on organizations and the natural environment, 1992-present: A review. Journal of Management 33: 637–664.

20.

Ferrell

Liang

Renneboog

(2016) Socially responsible firms. Journal of Financial Economics 122: 585–606.

21.

Flammer

Hong

Minor

(2019) Corporate governance and the rise of integrating corporate social responsibility criteria in executive compensation: Effectiveness and implications for firm outcomes. Strategic Management Journal 40: 1097–1122.

22.

Gaver

Ausin

(1995) Additional evidence on bonus plans and income management. Journal of Accounting and Economics 19: 3–28.

23.

Goldreyer

Diltz

(1999) The performance of socially responsible mutual funds: Incorporating sociopolitical information in portfolio selection. Managerial Finance 25: 23–36.

24.

Gomez-Mejia

Cruz

Berrone

(2011) The bind that ties: Socioemotional wealth preservation in family firms. Academy of Management Annals 5: 653–707.

25.

Gomez-Mejia

Haynes

Nunez-Nickel

, et al. (2007) Socioemotional wealth and business risks in family-controlled firms: Evidence from Spanish olive oil mills. Administrative Science Quarterly 52: 106–137.

26.

Gomez-Mejia

Makri

Lazzara-Kintana

(2010) Diversification decisions in family-controlled firms. Journal of Management Studies 47: 223–252.

27.

Gormley

Matsa

(2016) Playing it safe? Managerial preferences, risk, and agency conflicts. Journal of Financial Economics 122: 431–455.

28.

Griffin

Mahon

(1997) The corporate social performance and corporate financial performance debate: Twenty-five years of incomparable research. Business & Society 36: 5–31.

29.

Guenther

Hoppe

(2014) Merging limited perspectives: A synopsis of measurement approaches and theories of the relationship between corporate environmental and financial performance: Limited perspectives of the CEP-CFP link. Journal of Industrial Ecology 18: 689–707.

30.

Gul

Chen

Tsui

(2003) Discretionary accounting accruals, managers’ incentives, and audit fees. Contemporary Accounting Research 20: 441–464.

31.

Healy

Wahlen

(1999) A review of the earnings management literature and its implications for standard Setting. Accounting Horizons 13: 365–383.

32.

Hemingway

Maclagan

(2004) Managers’ personal values as drivers of corporate social responsibility. Journal of Business Ethics 50: 33–44.

33.

Holthausen

Larcker

Sloan

(1995) Annual bonus schemes and manipulation of earnings. Journal of Accounting and Economics 19: 29–74.

34.

Ioannou

Serafeim

(2015) The impact of corporate social responsibility on investment recommendations: Analysts’ perceptions and shifting institutional logics. Strategic Management Journal 36: 1053–1081.

35.

Jensen

(1986) Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review 76: 323–329.

36.

Jensen

Meckling

(1976) Theory of the firm: Managerial behavior, agency cost and ownership structure. Journal of Financial Economics 3: 305–360.

37.

Jones

(1995) Instrumental stakeholder theory: A synthesis of ethics and economics. Academy of Management Review 20: 404–437.

38.

Korea Chamber of Commerce and Industry (2006) A study on the status and implications of family firms in major countries. Policy study report, Korea Chamber of Commerce and Industry, Seoul, South Korea.

39.

Korea Fair Trade Commission (KFTC) (2015) 2014 Annual Report. Seoul, South Korea: KFTC.

40.

La Porta

Lopez-de-Silanes

Shleifer

(1999) Corporate ownership around the world. The Journal of Finance 54: 471–517.

41.

Lee

K-H

Cin

Lee

(2016) Environmental responsibility and firm performance: The application of an environmental, social and governance model. Business Strategy and the Environment 25: 40–53.

42.

Lee

K-H

Min

Yook

(2015) The impacts of carbon (CO2) emissions and environmental research and development (R&D) investment on firm performance. International Journal of Production Economics 167: 1–11.

43.

Lev

Petrovits

Radhakrishnan

(2010) Is doing good for you? How corporate charitable contributions enhance revenue growth. Strategic Management Journal 31: 182–200.

44.

Luffarelli

Awaysheh

(2017) The impact of indirect corporate social performance signals on firm value: Evidence from an event study. Corporate Social Responsibility and Environmental Management 25: 295–310.

45.

McConaughy

Walker

Henderson

, et al. (1998) Founding family controlled firms: Efficiency and value. Review of Financial Economics 7: 1–19.

46.

McConnell

Servaes

(1990) Additional evidence on equity ownership and corporate value. Journal of Financial Economics 27: 595–612.

47.

McGuire

Oehmichen

Wolff

(2019) Do contracts make them care? The impact of CEO compensation design on corporate social performance. Journal of Business Ethics 157: 375–390.

48.

McWilliams

Siegel

Wright

(2006) Guest editors’ introduction: Corporate social responsibility: Strategic implications. Journal of Management Studies 43: 1–18.

49.

Martínez

Stöhr

Quiroga

(2007) Family ownership and firm performance: Evidence from public companies in Chile. Family Business Review 20: 83–94.

50.

Martínez-Ferrero

Villarón-Peramato García-Sánchez

ÓIM

(2017) Can investors identify managerial discretion in corporate social responsibility practices? The moderate role of investor protection. Australian Accounting Review 27: 4–16.

51.

Maury

(2006) Corporate performance, corporate governance and top executive turnover in Finland. European Financial Management 12: 221–248.

52.

Morck

Shleifer

Vishny

(1988) Management ownership and market valuation: An empirical analysis. Journal of Financial Economics 20: 293–315.

53.

Orlitzky

Schmidt

Rynes

(2003) Corporate social and financial performance: A meta-analysis. Organization Studies 24: 403–441.

54.

Orlitzky

Siegel

Waldman

(2011) Strategic corporate social responsibility and environmental sustainability. Business and Society 50: 6–27.

55.

Ortiz-de-Mandojana

Bansal

(2016) The long-term benefits of organizational resilience through sustainable business practices. Strategic Management Journal 37: 1615–1631.

56.

Park

B-J

(2017) Auditors’ economic incentives and the sensitivity of managerial pay to accounting performance. Australian Accounting Review 83: 381–399.

57.

Pava

Krausz

(1996) The association between corporate social-responsibility and financial performance: The paradox of social cost. Journal of Business Ethics 15: 321–357.

58.

Peloza

(2006) Using corporate social responsibility as insurance for financial performance. California Management Review 48: 52–72.

59.

Prior

Surroca

Tribo

(2008) Are socially responsible managers really ethical? Exploring the relationship between earnings management and corporate social responsibility. Corporate Governance 16: 160–177.

60.

Reid

Toffel

(2009) Responding to public and private politics: Corporate disclosure of climate change strategies. Strategic Management Journal 30: 1157–1178.

61.

Roman

Hayibor

Agle

(1999) The relationship between social and financial performance: Repainting a portrait. Business & Society 38: 109–125.

62.

Shleifer

Vishny

(1989) Management entrenchment: The case of manager-specific investment. Journal of Financial Economics 25: 123–139.

63.

Smith

Amoako-Abu

(1999) Management succession and financial performance of family controlled firms. Journal of Corporate Finance 5: 341–368.

64.

Ullman

(1985) Data in search of a theory: A critical examination of the relationships among social performance, social disclosure, and economic performance of U.S. firms. Academy of Management Review 10: 540–557.

65.

Villalonga

Amit

(2006) How do family ownership, control and management affect firm value? Journal of Financial Economics 80: 385–417.

66.

Waddock

Graves

(1997) The corporate social performance-financial performance link. Strategic Management Journal 18: 303–319.

67.

Wang

Bansal

(2012) Social responsibility in new ventures: Profiting from a long-term orientation. Strategic Management Journal 33: 1135–1153.

68.

Weisbach

(1988) Outside directors and CEO turnover. Journal of Financial Economics 20: 431–460.

69.

Westhead

Cowling

(1998) Family firm research: The need for a methodological rethink. Entrepreneurship Theory and Practice 23: 31–56.

70.

Westhead

Coling

Storey

(2002) The scale and nature of family businesses. In: Carsrud

Brannback

(eds) Understanding the Small Family Business. New York: Springer, pp. 19–31.

71.

Wooldridge

(2016) Introductory Econometrics: A Modern Approach. 6th ed. Mason: South-Western Cengage Learning.