Institutional Background of Compensation Restriction

Since the Reform and Opening Up policy in 1978, the Chinese economy has been transforming from planned to market-oriented. Correspondingly, SOEs have experienced a series of critical reforms, including operational rights, control rights, and modernization of corporate governance (Lin, Lu, Zhang, & Zheng, Reference Lin, Lu, Zhang and Zheng2020; Stan, Peng, & Bruton, Reference Stan, Peng and Bruton2014). Nevertheless, SOEs are born with social functions (e.g., increasing employment, maintaining social stability and security, and providing various other social services to employees and communities) and economic functions. Although SOEs have experienced a range of privatization and modernization reforms, the state still retains control and substantial influence in resource allocations (Gu, Tang, & Wu, Reference Gu, Tang and Wu2020; Raynard, Lu, & Jing, Reference Raynard, Lu and Jing2020). With the state as the ultimate controlling owner, SOEs themselves reflect the will of the state, which tends to have a diverse set of social and political objectives (Bruton, Peng, Ahlstrom, Stan, & Xu, Reference Bruton, Peng, Ahlstrom, Stan and Xu2015; Stan et al., Reference Stan, Peng and Bruton2014). Meanwhile, retained state control over SOEs' top personnel, including compensation, evaluation, appointment, and removal, reinforces state interventions over SOEs' decision-making and operations (Du, Tang, & Young, Reference Du, Tang and Young2012; Li & Lu, Reference Li and Lu2020).

Facing severe inequality in society and increasingly growing executive compensation, as shown in Figure 1, the government prioritizes reducing income inequality and building a more harmonious society featured by common prosperity. To this end, the government selects SOEs as the exemplar, issuing the executives a tight compensation restriction (i.e., the Order). As a result, Figure 1 reveals a decrease in top executive compensation relative to their counterparts in private-owned enterprises. In addition to the political consideration, the Order shows some contradictions with conventional incentive designs (Murphy & Jensen, Reference Murphy and Jensen2018; Su et al., Reference Su, Jiang and Tian2020). Existing literature on aligning interests between the state and the executive focuses on two general themes for SOEs with multiple social and political functions (Bai et al., Reference Bai, Lu and Tao2006; Tirole, Reference Tirole1994). From the agency theory perspective, the first stream of literature focuses on direct economic incentives, such as compensation, in driving managerial efforts (Firth, Fung, & Rui, Reference Firth, Fung and Rui2006a; Groves, Hong, McMillan, & Naughton, Reference Groves, Hong, McMillan and Naughton1994). From the perspective of political influence, the second stream of literature focuses on political incentives, such as political career concerns, in driving managerial efforts (Cao et al., Reference Cao, Lemmon, Pan, Qian and Tian2019; Yang, Wang, & Nie, Reference Yang, Wang and Nie2013; Zheng, Li, Xu, Lin, & Zhao, Reference Zheng, Li, Xu, Lin and Zhao2012). Below, we briefly overview the institutional background and literature review on these two themes.

Figure 1.

The trend of executive compensation in three types of enterprises

Notes: TMT in the Y-axis denotes all the managers in management team, and the managers in the Y-axis denote all managers except for ordinary employees. For brevity, the numbers in the Y-axis denote the original value divided by 1,000 yuan.

Executive Incentives in SOEs

The classical theoretical perspective of executive compensation is normative agency theory, which predicts that compensation packages are an efficient design to incentivize executives (Conyon & He, Reference Conyon and He2011; Holmstrom & Milgrom, Reference Holmstrom and Milgrom1991; Jensen & Meckling, Reference Jensen and Meckling1976). In China, under the traditional central-planned economy during Chairman Mao's era, both managers (i.e., cadres) and employees were paid in terms of seniority and identity (i.e., worker/cadre) according to the national remuneration system based on collectivism and egalitarianism (Lin, Reference Lin2016; Zhu, De Cieri, & Dowling, Reference Zhu, De Cieri and Dowling1998). At that time, the pay aimed to cover an employee's daily consumption, and the pay gaps between ordinary employees and executives were marginal (Chizema, Liu, Lu, & Gao, Reference Chizema, Liu, Lu and Gao2015; Lin et al., Reference Lin, Cai and Li1998). During the 1980s, the contract responsibility system relaxed the compensation threshold, and some discretion over remuneration decisions was gradually granted to SOEs with total budget constraints. Adopting the contract responsibility system allows SOEs to retain part of profits, and employee and executive compensations could vary along with firm-level economic improvements. As the reform went further, market-based incentive schemes such as annual salary and bonuses were gradually adopted (Cao et al., Reference Cao, Lemmon, Pan, Qian and Tian2019; Groves et al., Reference Groves, Hong, McMillan and Naughton1994).

Correspondingly, executive compensation has changed significantly along with the reform of the state sector in recent decades (Chan & Ma, Reference Chan and Ma2011; Lin, Reference Lin2016). In the early period (1980–2003), the state primarily designed executive compensation through the contract responsibility system, including social and political objectives, in which profits were not the only objective (Bai & Xu, Reference Bai and Xu2005). Although the role of the state has been changing from the owner into a shareholder of SOEs, its retained control over top executives has enabled its continuous leverage over SOEs to realize social and political objectives (Hu & Xu, Reference Hu and Xu2022; Wang & Luo, Reference Wang and Luo2019).

After 2003, China established the State-owned Assets Supervision and Administration Commission (SASAC) to fulfill the interest of shareholders and directly monitor the SOEs, and performance-based incentives increased substantially (Conyon & He, Reference Conyon and He2011; Jiang & Kim, Reference Jiang and Kim2020). In general, the compensation of an SOE executive consists of three components: the base salary, performance-based bonuses, and mid-to-long-term incentive compensation. Then, SASAC takes economic performance and political and social indicators (e.g., employment, environmental protection, and social stability) into the evaluation system, and the executive compensations increase with productivity following a series of reforms (Du et al., Reference Du, Tang and Young2012; Mengistae & Xu, Reference Mengistae and Xu2004). Since the state tends to benchmark SOE executive pay with civil service pay, the actual pay-performance sensitivity in SOEs is usually weaker than in privately owned or foreign enterprises (Conyon & He, Reference Conyon and He2011; Lin, Reference Lin2016).

The multitasking nature of executive compensation of SOEs is well established, and many scholars investigate whether executive compensation matters (Bai et al., Reference Bai, Lu and Tao2006; Holmstrom & Milgrom, Reference Holmstrom and Milgrom1991). In general, the empirical findings reveal a positive relationship between executive pay and firm financial performance in SOEs, while this positive relationship is weaker when compared with the pay-for-performance link in private firms in China (Cao et al., Reference Cao, Lemmon, Pan, Qian and Tian2019; Conyon & He, Reference Conyon and He2011; Hu & Xu, Reference Hu and Xu2022; Mengistae & Xu, Reference Mengistae and Xu2004). In addition, a few studies confirm that executive compensation positively correlates with firm social performance in terms of job creation in Chinese SOEs (Gu et al., Reference Gu, Tang and Wu2020; Hu & Xu, Reference Hu and Xu2022). However, whether and why executive compensation contributes to a firm's overall CSR engagement and performance is largely underexamined.

Political Influences on SOEs

After decentralizing authority and privatizing SOEs in the last two decades, the government retained control over SOEs (Lin et al., Reference Lin, Lu, Zhang and Zheng2020; Sun & Tong, Reference Sun and Tong2003), and top executives in SOEs are quasi-officials in a political system and a labor market for professional managers (Hu & Xu, Reference Hu and Xu2022; Manion, Reference Manion1985). Thus, Chinese SOEs have two parallel personnel systems: regular corporate management and Party-state systems (Walder, Reference Walder1995; Wang, Yi, Zhang, & Peng, Reference Wang, Yi, Zhang and Peng2022). Cross-appointments between these two systems are institutionalized such that a corporate manager of a given rank holds a position of equivalent administrative rank in the Party-state system.Footnote ³ Leading personnel are moved back and forth between state bureaucracies and SOEs at the command of the Party-state, which provides SOE executives with political promotion motives regarding upward moves in the Party-state system (Wang et al., Reference Wang, Fu, Huang and Wang2014; Yang et al., Reference Yang, Wang and Nie2013).

Rooted in agency theory, dual-career tracks – business and politics – motivate SOE executives to balance multiple task requirements (Bai et al., Reference Bai, Lu and Tao2006; Bai & Xu, Reference Bai and Xu2005; Cao et al., Reference Cao, Lemmon, Pan, Qian and Tian2019). The political path refers to the job mobility track from an SOE executive position to a purely political position in the state system. In contrast, the business path refers to the job mobility track from one current occupational position to another within the business field (Lin et al., Reference Lin, Lu, Zhang and Zheng2020; Zhang, Liu, & Cai, Reference Zhang, Liu and Cai2015). Finally, the incentives in each track present a somewhat substituted relationship (Cao et al., Reference Cao, Lemmon, Pan, Qian and Tian2019; Chen, Guan, & Ke, Reference Chen, Guan and Ke2013). For example, Cao et al. (Reference Cao, Lemmon, Pan, Qian and Tian2019) investigated the relationship between political promotion and firm performance of SOEs. The results show that executive pay-for-performance sensitivity is higher when SOE executives are less likely to receive a political promotion. The motives for pursuing job mobility in the political job market could substitute explicit compensation incentives.

Political loyalty and conformity with political responsibility are essential for career advancement in political screening and incorporation (Walder, Reference Walder1995). Facing multiple demands from the government simultaneously, the priority of resource allocation will vary with urgency and temporal emphasis. Under the state's priority over sustainable development, social welfare expansion, and the creation of a ‘people-oriented’ harmonious society, achievements in social policy areas such as employee treatment, public security, resource consumption, resource conservation, and environmental protection became new imperatives for political promotions (Li & Lu, Reference Li and Lu2020; Liu & Xiao, Reference Liu and Xiao2015). For example, environmental protection performance carries more weight in the evaluation criteria of officials than in the past.Footnote ⁴ Accordingly, executives must direct their efforts to social policy areas besides financial performance to secure a political promotion.

The SOE executives face two intertwined channels with distinct demands ranging from economic and social performance. Research emphasizes the political influence on SOEs' CSR engagement and performance, while the agents executing social obligations should not be negligible. In the meantime, research deriving from agency theory always focuses on the interest conflicts between state shareholders with a social orientation and private shareholders with a profit maximization orientation. We need more details about the role of specific incentive tools (e.g., compensation) in the executive action of CSR, especially the managerial attention to incentive changes.

Effect of Compensation Restriction on SOEs’ CSR Performance

In this study, we are interested in the SOE's social performance in response to the compensation restriction policy initiated by the government. According to the prior literature, two interrelated mechanisms lead to the policy effect: firstly, the political influence stemming from government oversight; and secondly, economic incentives that foster alignment of interests between the principal and agent.

Disciplinary Pressure of Performance Declines

Performance declines impose disciplinary pressures on poorly performing executives, and the subsequent organizational responses are contingent upon the executives' subjective interpretations and cognitive appraisals of the situation (Audia, Locke, & Smith, Reference Audia, Locke and Smith2000; Firth, Fung, & Rui, Reference Firth, Fung and Rui2006b; Jordan & Audia, Reference Jordan and Audia2012). In this light, the shortfalls of economic performance will threaten the business prospect and manager's career, and hence lead to conserving resources through cost-cutting and efforts for greater economic efficiency and performance recovery (Staw, Sandelands, & Dutton, Reference Staw, Sandelands and Dutton1981; You, Shin, & Chung, Reference You, Shin and Chung2023; Zhong, Ren, & Song, Reference Zhong, Ren and Song2022). For corporate social performance, constant resource investment is necessary to maintain the status quo of corporate social performance (Wang & Choi, Reference Wang and Choi2013). When the firm faces earnings pressure, which requires executives to protect and prioritize economic returns, the manager responds by reducing the likelihood and performance of corporate social engagement (Zhang, Zhang, Jia, & Ren, Reference Zhang, Zhang, Jia and Ren2020). Hence, the disciplinary pressures ensuing from diminished performance may exacerbate the demotivational effects on executives, potentially disincentivizing them from prioritizing social performance in their strategic decision-making processes.

Compensation arrangements aim to align the interests of executives with the interest of shareholders, and the performance status quo determines the alignment effect (Hambrick, Reference Hambrick2007; Popli & Raithatha, Reference Popli and Raithatha2023). A performance decline situation could shape how executives view and respond to environmental changes and affect firms' decisions by filtering and interpreting the ambient situations (Campbell, Bilgili, Crossland, & Ajay, Reference Campbell, Bilgili, Crossland and Ajay2023; Hambrick, Finkelstein, & Mooney, Reference Hambrick, Finkelstein and Mooney2005). For example, decreasing financial resources resulting from substantial performance declines affect the emphasis on resource-freeing rather than resource-consuming organizational change in response to performance shortfalls (Kuusela et al., Reference Kuusela, Keil and Maula2017; Mannor, Wowak, Bartkus, & Gomez-Mejia, Reference Mannor, Wowak, Bartkus and Gomez-Mejia2016). Thus, when executives experience a ceiling on their pay regardless of how hard they work, especially performance declines that lead to increasing resource constraints, this would further demotivate their engagement in long-term social issues.

With respect to the disincentive effect of compensation restriction, disciplinary pressures and resource constraints resulting from performance declines will further urge the executives to assess the prospects concerning business tracks and then conserve resources and preserve the firm's economic trajectory. Therefore, we integrate the previous ideas from performance disciplinary and lose aversion will operate in tandem to influence executive decisions and actions regarding social domains, the executives experiencing performance declines might be more sensitive to the change in compensation arrangements. By contrast, a lower performance decline could leave room for executives to conserve resources and function as a financial buffer that gives executives the freedom to respond to compensation restrictions by deviation from long-term social performance. As such, we posit that top executives in a situation of performance decline will further attenuate their engagement in social issues. Therefore, we propose the following:

Moderating Role of Executive's Relative Pay Gap

Perceptions of financial wealth relative to salient referents drive executive behavior, and executive compensation research shows that the effectiveness of pay arrangements in directing managerial efforts is partially determined by executives' perception of fairness or equality about their pay (Bosse & Phillips, Reference Bosse and Phillips2016; Devers, Cannella Jr, Reilly, & Yoder, Reference Devers, Cannella, Reilly and Yoder2007). In a social context, almost everyone cares about fairness to some degree and is more sensitive to underpayment than overpayment (Adams, Reference Adams and Berckowitz1965; Fong, Reference Fong2010). Empirical evidence confirms that a negative violation of the fairness norm (i.e., being unfairly underpaid) will lower the valence of monetary incentives due to the rising concern about fairness (Shaw, Reference Shaw2014; Wade et al., Reference Wade, O'Reilly and Pollock2006). We posit that the observed divergence in executive responses to compensation restrictions stems from a complex interplay of individual and contextual factors, leading to a heterogeneous array of strategic reactions among executives. As perceptions of unfairness are determined through social comparison with other similarly situated executives (Wade et al., Reference Wade, O'Reilly and Pollock2006), we introduce the executive's relative pay gap as a moderator to the main relationship.

Executives are sensitive to underpayment, and a large relative pay gap will induce their perceptions of unfairness (Coles et al., Reference Coles, Li and Wang2018; Fong, Reference Fong2010; Fong, Misangyi, & Tosi, Reference Fong, Misangyi and Tosi2010). This perception initiates a recalibration of the executives' economic calculus, subsequently amplifying their propensity to mitigate their efforts in the face of imposed compensation limitations. Executives who perceive their compensation as inadequate may seek to rectify this discrepancy by either augmenting the output through enhanced performance outcomes, such as securing a salary increment, or by adjusting the input by moderating their exerted efforts (Fong et al., Reference Fong, Misangyi and Tosi2010). We contend that executives exhibit a pronounced preference for the latter course of action, given the diminished agency in influencing outcomes within a framework of compensation constraints, thereby necessitating a strategic reallocation of their efforts (Coles et al., Reference Coles, Li and Wang2018; Cowherd & Levine, Reference Cowherd and Levine1992).

Accordingly, the imposition of compensation restrictions, particularly when executives are confronted with substantial relative pay disparities, may precipitate a reduction in their investment of effort toward CSR initiatives, culminating in a significant degradation of CSR performance metrics. In contrast, a small relative pay gap is unlikely to induce perceptions of unfairness. Because the norm is that SOE executives are state officials receiving civil service pay – which should be lower than most top executives across corporations (Chen et al., Reference Chen, Guan and Ke2013; Lin, Reference Lin2017). Thus, we do not expect a small relative pay gap to influence an executive's response to compensation restriction. Accordingly, we should observe more CSR performance declines among SOEs whose top executives experience large relative pay gaps. This leads to the following hypothesis:

Moderating Role of Executive Political Career Horizon

Whereas compensation restriction disincentivizes managerial efforts in social areas, the likelihood of promotion in the political ladder may somewhat recalibrate the managerial interpretation of compensation restriction (Cao et al., Reference Cao, Lemmon, Pan, Qian and Tian2019; Hsu et al., Reference Hsu, Liang and Matos2023). Immediate monetary incentives and future career opportunities are the two main components of career concerns that motivate executives' efforts simultaneously (Cao et al., Reference Cao, Lemmon, Pan, Qian and Tian2019; Holmström, Reference Holmström1999). For top executives of SOEs, the future promotion to administrative rank is a kind of deferred compensation. Executives pursuing political careers tend to direct their efforts toward firm political performance and be less sensitive to current economic incentives (Hu & Xu, Reference Hu and Xu2022; Zheng et al., Reference Zheng, Li, Xu, Lin and Zhao2012). Social and environmental issues have become a priority in the state's goal (Zuo, Reference Zuo2015). To secure a political promotion, these executives must effectively signal their compliance with state demands and competence in fulfilling political objectives set by the state (Du et al., Reference Du, Tang and Young2012; Liu & Xiao, Reference Liu and Xiao2015; Zheng et al., Reference Zheng, Li, Xu, Lin and Zhao2012). Thus, top executives subjected to the compensation restriction may continue advancing the firm's political performance due to their political promotion motives.

Executive political career horizon refers to the possibility of future mobility to a higher position in the political hierarchy (Li & Lu, Reference Li and Lu2020; Wang & Luo, Reference Wang and Luo2019). Under the condition where a maximum compensation threshold is established, executives with a high likelihood of political promotion are less sensitive to compensation restriction and more likely to continue to dedicate their efforts to improve firm CSR performance when their compensation peaks. Prior studies confirm that strong political promotion incentives can substitute for direct monetary incentives in aligning SOE executives' interests with that of the state (Cao et al., Reference Cao, Lemmon, Pan, Qian and Tian2019; Hu & Xu, Reference Hu and Xu2022). Thus, immediate compensation restriction might reduce managerial efforts on CSR improvement. However, executives with extended career horizons could experience a mitigated demotivational effect, attributed to their heightened anticipation of future promotional opportunities within the organizational hierarchy.

In contrast, when executives approach the culmination of their careers with a diminished prospect of political advancement, they exhibit a reduced propensity to sustain their efforts toward enhancing firm CSR performance when their compensation peaks. In such instances, these executives are inclined to preserve the established baseline level of CSR performance, driven by a strategic imperative to avert potential demotion or disciplinary measures (Liu & Xiao, Reference Liu and Xiao2015). In this light, SOE executives close to retirement age have limited political promotion opportunities; they are less likely to devote efforts to performance targets that contribute to career advancement (Li & Lu, Reference Li and Lu2020; Wang, Zhu, Chen, & Luo, Reference Wang, Zhu, Chen and Luo2021). Hence, we argue that executives with long political career horizons, as compared to short political career horizons, will be less sensitive to current monetary rewards and demotivated by compensation restrictions. Accordingly, we propose that firms whose executives are subject to compensation restriction are less likely to experience CSR performance decline when their executives have a long political career horizon. Therefore, we propose the following:

Moderating Role of Firm Exposure to Socialist Imprint

Firms' founding era shapes historical and ideological underpinnings for organizational culture, which continue to permeate individual cognition and behavior in reinforcing the firm's imprint on individual decision-making. Given that ‘organizations formed at one time typically have a different social structure from those formed at another time’ (Stinchcombe, Reference Stinchcombe and March1965: 154), the persistent effect of imprinting could influence current decision-making (Marquis & Tilcsik, Reference Marquis and Tilcsik2013; Stinchcombe, Reference Stinchcombe and March1965). In particular, the government accords precedence to social stability, thereby influencing firms to align their operational behaviors with the foundational imprints established by initial policies (Kogut & Zander, Reference Kogut and Zander2000; Lin et al., Reference Lin, Cai and Li1998; Marquis & Qian, Reference Marquis and Qian2014). In the context of SOEs, executives are subject to the enduring influence of the socialist imprint, which serves as a critical information filter. This filter channels their focus toward fulfilling political mandates across a spectrum of social concerns, transcending the pursuit of mere economic incentives. The socialist imprinting significantly molds firm-specific capabilities, exerting a pervasive influence that extends to both individual and organizational behaviors, particularly among those who have experienced the era of Mao (Han & Zheng, Reference Han and Zheng2016; Raynard et al., Reference Raynard, Lu and Jing2020; Wang & Luo, Reference Wang and Luo2019). In this context, the imprints of collectivism and egalitarianism, deeply ingrained within organizational structures, practices, and operational routines, manifest with particular intensity in SOEs established during Chairman Mao's era, which is due to the intensive training and socialization processes of the party's doctrines and values (Raynard et al., Reference Raynard, Lu and Jing2020; Zhang, Ren, et al., Reference Zhang, Ren and Wu2023). Consequently, as the economic landscape underwent a paradigm shift toward a market-oriented economy, enterprises established during the era of planned economy exhibited a more pronounced imprint of socialist values, reflecting the ideological legacy of their foundational period (Boeker, Reference Boeker1989; Kriauciunas & Kale, Reference Kriauciunas and Kale2006; Wang & Luo, Reference Wang and Luo2019; Zhang, Ren, et al., Reference Zhang, Ren and Wu2023).

During the socialist training and socialization processes, which historically emphasized people-oriented and egalitarian preferences (Haack & Sieweke, Reference Haack and Sieweke2018; Han & Zheng, Reference Han and Zheng2016), these imprints are likely to persist and exert a corrective influence on current decision-making, particularly in the domain of compensation policies concerning compensation restriction. Although socialist imprints will decay during the market transition (Haack & Sieweke, Reference Haack and Sieweke2018; Zhang, Ren, et al., Reference Zhang, Ren and Wu2023), ‘training programs and wage systems are more internal and people-oriented’ (Kriauciunas & Kale, Reference Kriauciunas and Kale2006: 674), which suggests that these imprints continue to channel executive focus toward the promotion of social equality within compensation structures. We posit that executives operating within SOEs that bear the hallmark of socialist imprinting are inclined to integrate collective and egalitarian values into their decision-making frameworks. Consequently, the motivational potency of monetary incentives is recalibrated, influencing their directional focus toward the adoption of compensation restrictions.

Drawing from the preceding arguments, enterprises that were established during the period when socialist values were deeply entrenched in China's economic paradigm were inherently shaped by a social structure and institutional environment that championed labor protection and a human-centric ethos. Meanwhile, SOEs emerged as extensions of the state, embodying the principles of collectivism and egalitarianism, while monetary incentives were strongly disdained with a strong emphasis on ‘spiritual encouragements’ (Raynard et al., Reference Raynard, Lu and Jing2020). Considering the profound socialization exerted by the Party during the socialist era, we propose that SOEs established during this period are likely to be imbued with political values that advocate collectivism and egalitarianism. Consequently, top executive in such firms, underpinned by collective values, are inclined to prioritize communal interests over individualistic pursuits. Furthermore, in decision-making processes, top executives in those organizations that embrace egalitarian principles are more likely to favor compensation structures characterized by minimal pay disparities between executives and their rank-and-file employees (Haack & Sieweke, Reference Haack and Sieweke2018; Kriauciunas & Kale, Reference Kriauciunas and Kale2006). Accordingly, organizations with a more pronounced socialist imprint are hypothesized to exhibit a diminished sensitivity to the disincentivizing effects of compensation restrictions, particularly when compared with firms lacking such historical influences. This diminished sensitivity arises because compensation restrictions align with the promotion of social equity, the reduction of income disparities, and the core values of these organizations. Executives within such firms are likely to experience a reduced motivational impact when faced with compensation limitations, as these restrictions resonate with their intrinsic egalitarian ideals permeated through cognition and socialization. In this context, the alignment of organizational policies with executives' values can partially attenuate the disincentive effect of compensation restrictions. This attenuation occurs not by diminishing the executives' efforts but by redefining the significance of monetary incentives in relation to their overall value system. We contend that social imprints could significantly rectify the attention toward income equality and attenuate the decrease in CSR performance. Therefore, we propose the following hypothesis:

To summarize the above hypotheses, the general framework of the study is shown in Figure 2.

Figure 2.

Conceptual model

Archival Methods

Estimation strategy

We employ a standard DID approach to examine our model. H1 is tested based on the following model specification:

(1)$$\eqalign{{\rm CS}{\rm R}_{{\rm tota}{\rm l}_{i, t}}{\rm} = \beta _0{\rm} + \beta _1{\rm Trea}{\rm t}_i{\rm} + \beta _2{\rm Pos}{\rm t}_t{\rm} + \beta _3{\rm Trea}{\rm t}_i \times {\rm Pos}{\rm t}_t \cr {\rm} + \beta _i{\rm Contro}{\rm l}_{i, t-1}{\rm} + {\rm Firm}\,{\rm FE} + {\rm Year}\,{\rm FE} + \varepsilon _{i, t}} $$

In this equation, i indicates firm, t indicates the observation year, and $\varepsilon _{i, t}\;$ is the error term. ${\rm CS}{\rm R}_{{\rm tota}{\rm l}_{i, t}}$ indicates the CSR performance of firm i in year t. Treat_i indicates whether top executives of firm i are subjected to compensation restriction. Post_t indicates whether year t is in the post-event period. The coefficient of Treat_i × Post_t β ₃ gives the estimates of the main effect, which is our focus. Control_i,t−1 represents the matrix of control variables in the last period. Firm FE and Year FE represents firm-specific fixed effects and year-specific fixed effects, respectively.

The moderating effect is tested by adding the moderator and the interaction terms into equation (1). We test H2–5 separately, based on the following model specification:

(2)$$\eqalignno{{\rm CS}{\rm R}_{{\rm tota}{\rm l}_{i, t}}{\rm} & = \beta _0{\rm} + \beta _1{\rm Trea}{\rm t}_i{\rm} + \beta _2{\rm Pos}{\rm t}_t{\rm} + \beta _3{\rm Trea}{\rm t}_{\rm i}\times {\rm Pos}{\rm t}_t \cr & \quad+ \beta _4{\rm Trea}{\rm t}_i\times {\rm Pos}{\rm t}_t\times {\rm Moderato}{\rm r}_i{\rm} + \beta _5{\rm Moderato}{\rm r}_i \cr & \quad+ \beta _6{\rm Moderato}{\rm r}_i\times {\rm Pos}{\rm t}_t{\rm} + \beta _7{\rm Moderato}{\rm r}_i\times {\rm Trea}{\rm t}_i \cr & \quad+ \beta _i{\rm Contro}{\rm l}_{i, t-1}{\rm} + {\rm Firm}\,{\rm FE} + {\rm Year}\,{\rm FE} + \varepsilon _{i, t}} $$

In this equation, we add the Moderator_i and three interaction terms, including Moderator_i × Post_t, Moderator_i × Treat_i, and Treat_i × Post_t × Moderator_i, based on equation (1). The coefficient of the last three-way interaction term β ₄ gives the estimates of the moderating effect, which is our focus. It is noteworthy that we report regression results with two different treatment groups. Regressions under the subtitle of ‘Central SOEs’ use Central SOEs as the treatment group, with local SOEs and private firms as the control group. Regressions under the subtitle of ‘All SOEs’ use both Central and local SOEs as the treatment group, with private firms as the control group.

Robustness Checks

To test the robustness of the regression results, we perform a battery of robustness checks, including the parallel trend assumption for the validity of a quasi-experimental analysis, alternative raters for CSR measure, and alternative time windows and samples.

Graphical checks on parallel trend assumption

We begin the robustness check with validation tests of our quasi-experiment design. In this section, to verify the parallel trend assumption, we examine whether the Order decreases the CSR performance of SOEs and whether there is a significant difference between SOEs and other enterprises before the Order implementation. Following previous research (Fiechter, Hitz, & Lehmann, Reference Fiechter, Hitz and Lehmann2022), we plot the gradual decrease in CSR performance after the Order. In Figure 3, the coefficient differences prior to the Order, the Central SOEs, and other firms are associated with parallel trends. Unfortunately, the difference pattern between all SOEs and other firms is somewhat confusing. The reason might be that the Central SOEs are the direct targets of the Order, while the other SOEs only have a somewhat volatile spillover effect. Given that we consider Central SOEs and all SOEs in the regressions, the results show a similar pattern, and we can still believe the empirical findings.

Figure 3.

Dynamics of the Order and CSR performance difference

Notes: (a) The coefficient change in CSR performance between Central SOEs and other firms. (b) The percentage change in CSR performance between all SOEs and other firms. A six-year window spanning from three years before to three years after the Order enactment. The dotted lines represent the 99% confidence intervals of the estimated effect based on standard errors clustered at the firm level. The year of the Order enactment serves as the benchmark year and is omitted from the presentation.

Alternative raters of social performance to mitigate the concern of certain rater biases

To mitigate the concern of potential rater biases in our dependent variable, we use an alternative measure of CSR performance. Several studies have questioned the agreement across different raters of CSR and suggest we be cautious in interpreting the results of a specific rater (Berg, Koelbel, & Rigobon, Reference Berg, Koelbel and Rigobon2022; Chatterji, Durand, Levine, & Touboul, Reference Chatterji, Durand, Levine and Touboul2016). To mitigate this concern, we use firms' CSR data from the China Listed Firm Corporate Social Responsibility Research Database maintained by the GTA (Zhang, Wang, et al., Reference Zhang, Wang and Zhou2020; Zhang, Zhou, et al., Reference Zhang, Zhou and Lyles2023) and re-run our analyses in terms of nine fields except for deficiency field. Regression results are reported in Table 7. As shown in Table 7, the coefficient of Treat × Post in model 2 is significantly negative (β = −0.575, p < 0.01), indicating that compensation restriction leads to significant CSR performance declines among Central SOEs. Likewise, the coefficient of Treat × Post in model 4 is significantly negative (β = −0.504, p < 0.01), confirming that compensation restriction leads to significant CSR performance declines among all SOEs. These results provide consistent support for the hypothesized negative effect of compensation restriction on CSR performance, suggesting that the concern of potential rater biases is not an issue in our study.

Table 7.

Alternative raters of CSR performance

Notes: The t statistics reported in parentheses are based on all standard errors clustered at the firm level. All tests are two-tailed. ***p < 0.01, **p < 0.05, *p < 0.1.

Sample range changes to mitigate omission and arbitrary deletions of the observations

To mitigate the concern of omissions and arbitrary deletions of the observations, we conduct three additional analyses in which we change the sample range. Regression results are reported in Table 8. As the Order was announced in 2014, firms that do not experience its announcement and formal enforcement will lack comparisons. To mitigate this concern, we exclude firms established before 2014 from our final sample, and then we re-run our analysis with all SOEs as the treatment group and the rest of the private firms as the control group. As shown in model 1 of Table 8, the coefficient of Treat × Post is negative and significant (β = −3.345, p < 0.01), which is consistent with the prediction of H1. We also exclude firms that went IPO before 2014 from our final sample, and the results provide consistent support for H1; the coefficient of Treat × Post in model 2 is negative and significant (β = −3.470, p < 0.01). Finally, we include financial firms in our final sample, the results providing further support for our theory; the coefficient of Treat × Post in model 3 is negative and significant (β = −3.328, p < 0.01). Together, these results suggest that omissions are not an issue in our study.

Table 8.

Sample changes in CSR performance

Notes: The t statistics reported in parentheses are based on all standard errors clustered at the firm level. All tests are two-tailed. ***p < 0.01, **p < 0.05, *p < 0.1.

Alternative observation windows to validate the timing of policy effect

Following previous research (Shi, Xia, & Meyer-Doyle, Reference Shi, Xia and Meyer-Doyle2022), we adopt three alternative time windows (i.e., 2009–2019, 2011–2017, and 2013–2016) to verify the negative effect of compensation restriction on CSR performance. Regression results are reported compared with the period of 2010–2018 used in our main analyses (see Table 3). As shown in Table 9, using both Central SOEs and local SOEs as the treatment group, the coefficient of Treat × Post across models 2, 3, and 4 is negative and significant at p < 0.1, suggesting that compensation restriction leads to significant CSR performance declines among SOEs. These results provide further support for H1.

Table 9.

Alternative time windows of CSR performance

Notes: The t statistics reported in parentheses are based on all standard errors clustered at the firm level. All tests are two-tailed. ***p < 0.01, **p < 0.05, *p < 0.1.

Further Analyses

To further explore the heterogeneous effect of compensation restriction on CSR components, we test the relationship between compensation restriction on subdimensions of CSR. Moreover, the initial purpose of the Order is to alleviate inequality and boost common prosperity, and we further test the policy effect on compensation changes of top executives and all employees.

CSR subdimensions

The above results and arguments show that the compensation restriction disincentivizes the top executives in overall CSR performance, while we still do not know the effect on specific dimensions of CSR. CSR is a kind of complex corporate activity comprising multiple dimensions ranging from the inside employee to the outside society of the firms (Farooq, Rupp, & Farooq, Reference Farooq, Rupp and Farooq2017; Hawn & Ioannou, Reference Hawn and Ioannou2016; Hsu et al., Reference Hsu, Liang and Matos2023). To elucidate the underpinnings of the nexus between compensation constraints and the firm's social responsibility decision-making, our study conducts a meticulous analysis of the impact of compensation restrictions across the various subdimensions of social responsibility. The empirical findings reveal that the imposition of compensation restrictions exerts a significant negative influence on the investment allocated to employee welfare and the cultivation of client and customer relationships. Furthermore, there is a marked decline in the firm's environmental performance metrics, indicating a broader impact on the company's social responsibility initiatives. However, SOEs enhance the responsibility of the shareholder – the government, which dominates and directs the ultimate resource allocation. The reason might be that financial performance is a critical criterion of performance evaluation for SOEs (Du et al., Reference Du, Tang and Young2012; Zuo, Reference Zuo2015). Noticing that the performance in society dimension does not significantly change after the enforcement of compensation restriction, the reason might be that the SOEs are an extension of the government and are responsible for societal issues (Bai et al., Reference Bai, Lu and Tao2006; Lin et al., Reference Lin, Cai and Li1998). Accordingly, the compensation restriction will not change the investment in societal issues. Considering each subdimension in isolation, one with a more accurately observed outcome would have a higher-powered incentive because the outcome is easy to motivate (Flammer et al., Reference Flammer, Hong and Minor2019). Overall, the total effect (TE) of compensation restriction on CSR is negative, with the specific effect on different dimensions depending on the degree of priority of SOEs' goals (Table 10).

Table 10.

Compensation restriction and CSR subdimensions

Notes: The t statistics reported in parentheses are based on all standard errors clustered at the firm level. All tests are two-tailed. ***p < 0.01, **p < 0.05, *p < 0.1.

Top executive compensation change

Although Figure 1 shows top executives' compensation change before and after the Order, we still know little about the effect except for the role of firm factors in this process. We employ the regression approach to identify the net effect of the policy with a series of firm characteristics and fixed effects. Apart from the allowance, top executive compensation has a significant reduction after the Order, which verifies that the policy restricts the final compensation substantially (Tables 11 and 12).

Table 11.

Top executive compensation changes

Notes: All compensation-related values are divided by 10,000. The t statistics reported in parentheses are based on all standard errors clustered at the firm level. All tests are two-tailed. ***p < 0.01, **p < 0.05, *p < 0.1.

Table 12.

Compensation restriction and employee compensation

Notes: All compensation-related values are divided by 10,000. The t statistics reported in parentheses are based on all standard errors clustered at the firm level. All tests are two-tailed. ***p < 0.01, **p < 0.05, *p < 0.1.

Basic Procedure

We conducted an experimental vignette consistent with prior studies, as shown in the Appendix (Alesina, Ferroni, & Stantcheva, Reference Alesina, Ferroni and Stantcheva2021). To balance the internal and external validity, the participants were MBA students (Audia et al., Reference Audia, Locke and Smith2000; Krause, Whitler, & Semadeni, Reference Krause, Whitler and Semadeni2014). As is common in experimental studies, participants were told that their current role was the top executive of a central SOE, and then we randomly assigned subjects to a hypothetical SOE executive that either met or did not meet a compensation restriction situation.Footnote ¹⁰ To this end, we provided identical information about the policy background and the firm's business scope to all participants, and the only difference was the trend of compensation change in the following years. In doing so, the descriptive scenarios could differentiate the compensation restriction from the consistent compensation policy, and the observed behavioral differences reflect the compensation change. Finally, the sample contained 79 participants, consisting of 42 participants in the treatment group and 37 participants in the control group.

Previous research suggests that subjective assessment effectively evaluates CSR performance (Latif & Sajjad, Reference Latif and Sajjad2018; Ong, Mayer, Tost, & Wellman, Reference Ong, Mayer, Tost and Wellman2018). Accordingly, we use Ong et al. (Reference Ong, Mayer, Tost and Wellman2018) eight-item scale to capture the executive engagement in social issues, such as ‘gives adequate contributions to charities’ and ‘cares about the well-being of people in general (i.e., the well-being of employees, customers, suppliers, shareholders, local community)’.

Moreover, given the reduction of engagement in the social area we observed in the compensation restriction situation, we next use data from the self-reported survey to measure the potential mechanisms driving the effect. We explore the conceptual predictions that existing literature suggests reactions following the compensation change, which, in turn, could lead to performance change in social areas. On the one hand, based on economic calculation, the work attitude change in response to compensation change is located in work engagement and work withdrawal (Kong, Park, & Peng, Reference Kong, Park and Peng2023). For all survey questions, we ask respondents to agree or disagree with the proposed statements, using a standard 6-point Likert scale ranging from Strongly Disagree (=1) to Strongly Agree (=6).

Work engagement refers to individuals' psychological connection with their work role and pertains to their full investment of physical, emotional, and cognitive resources in work (Christian, Garza, & Slaughter, Reference Christian, Garza and Slaughter2011; Lin, Reference Lin2010; Schaufeli, Bakker, & Salanova, Reference Schaufeli, Bakker and Salanova2006). In this study, we use Lin's (Reference Lin2010) six-item scale to measure work engagement, such as ‘At my work, I feel full of energy’ and ‘I am immersed in my work’. Similarly, work withdrawal refers to individuals' disengagement from their work role to alleviate anxiety and escape the stressful work environment (Kong et al., Reference Kong, Park and Peng2023; Scott & Barnes, Reference Scott and Barnes2011). In this study, we use Kong et al.'s (Reference Kong, Park and Peng2023) four-item scale to measure work withdrawal, such as ‘Came to work late without permission’ and ‘Taken a longer break than you were allowed to take’.

On the other hand, we consider the change in managerial interpretation of perceived social pressure stemming from political view in the process. In this light, perceived social pressure refers to managers' perceptions of institutional pressures pertaining to their focus on proactive social engagement (Buysse & Verbeke, Reference Buysse and Verbeke2003; Yang, Wang, Zhou, & Jiang, Reference Yang, Wang, Zhou and Jiang2019). In this study, we use Yang et al.'s (Reference Yang, Wang, Zhou and Jiang2019) four-item scale to measure perceived social pressure, such as ‘National (regional) governments influence our environmental issues management’ and ‘Local public agencies influence our environmental issues management’.

Effects of the intervention on social performance and exploration of potential mechanisms

Test for the proposed main relationship

Based on the scenario description, we analyzed the difference in mean value in these two groups, as is common in prior studies (Crilly, Ni, & Jiang, Reference Crilly, Ni and Jiang2016; Exley, Reference Exley2018). Figure 4a reveals the significant difference in corporate social engagement between the compensation restriction and consistent compensation groups. When the executives experience a compensation restriction, the participants have less active engagement (mean = 3.61) in social issues than their counterparts (mean = 4.50; p < 0.01).

Figure 4.

Level choosing to engage in social areas

Notes: This graph displays the level of individuals' motivation to engage in the corporate socially responsible areas across the treatment and control groups.

In addition, we also examined the relationship between compensation restriction and potential reactions from the executives. In response to the compensation restriction, managerial cognition would change around the policy. With respect to work engagement and work withdrawal, the participants showed a lower level of engagement in the compensation restriction group (mean = 3.56) than their counterparts (mean = 4.23; p < 0.01). By contrast, the work withdrawal level did not show a very significant difference (means = 2.29 vs 1.82, p = 0.06). Meanwhile, we also compare the perception difference about social pressure; the participants showed a lower level of social pressure in the compensation restriction group (mean = 3.67) than their counterparts (mean = 4.12; p < 0.03). Figures 4b–4d depict these results.

Test for underlying mechanisms

To examine the proposed mechanisms, we employ the conventional three-step method for testing mediation consistent with prior experimental studies (Exley, Reference Exley2018; Flammer & Kacperczyk, Reference Flammer and Kacperczyk2019; Krause et al., Reference Krause, Whitler and Semadeni2014). Models 1–4 in Table 13 further support the above independent sample t-test. Model 5 in Table 13 suggests that work engagement mediates the negative relationship between compensation restriction and corporate social performance. Moreover, the coefficient of perceived social pressure in Model 6 in Table 13 is statistically significant, suggesting that social pressure is also a force driving the reduction of corporate social performance.

Table 13.

Compensation restriction and level choosing to engage in social performance

Notes: The t statistics reported in parentheses. All tests are two-tailed. ***p < 0.01, **p < 0.05, *p < 0.1.

The above results provide a preliminary portrait of the mediating effect, and we use sem command in Stata 17 to further examine the indirect effect (IE), direct effect (DE), and TE of the mediation with structural equation model consistent with recent research (Portocarrero & Burbano, Reference Portocarrero and Burbano2024). The results in Table 14 support a mediating effect of work engagement reduction (IE = −0.183, p < 0.1) and perceived social pressure (IE = −0.144, p < 0.1), and the work withdrawal (IE = −0.032, p > 0.1) does not mediate the process. In sum, the reaction of managerial cognition in response to compensation restriction demotivates the work engagement and perception of social pressure, leaving the work withdrawal alone. This pattern may be because top executives would not shirk basic responsibility despite dissatisfaction with compensation arrangements.

Table 14.

Results of simple mediation analyses

Notes: Standard error in parentheses. Observation number is 79. ***p < 0.01, **p < 0.05, *p < 0.1.

In conclusion, the results of our experiment supplemented and extended the conclusions from archival study, which further supported the negative effect of compensation restriction on corporate social performance and uncovered the economic and political forces underlying the negative relationship. It offers evidence for the causal inference and disincentive effect that explains this pattern.