Catch One and Lose Another? Executive Compensation Restriction and Corporate Social Responsibility in State-Owned Enterprises

Abstract

This study examines how the managerial interpretation of incentive arrangement affects corporate engagement in social areas, as reflected in corporate social performance, from two interrelated perspectives: the political influence view and the normative agency view. Building the theoretical framework on state-owned enterprise (SOE) executives' dual-career tracks perspective, we contend that economic factors (performance decline and relative pay gap) and political factors (socialist imprints and political career horizon) could divergently reshape the interpretation of incentive arrangement on corporate social performance. Using ‘Pay Ceiling Order’ as a quasi-natural experiment context, a secondary analysis, and a controlled experiment reveal that compensation restriction on top executives causes a decrease in corporate social performance. This relationship is weakened when there are stronger socialist imprints inherited by a focal firm and when the executives have a longer political prospect. In contrast, the relationship is strengthened when firms face severe performance declines and when the executives' compensation is relatively lower than peers. The findings show that compensation is an indispensable incentive joining with political and economic factors, enabling SOEs to engage in social areas. We discuss the implications of understanding top executive incentives with incentive arrangements and how the government purpose influences top executive responses to compensation incentive in ways that matter for long-term social value.

摘要

本研究从两个相互关联的视角：政治影响观和规范代理观，考察了管理层对激励方案解释如何影响企业在社会领域的参与，以及这种参与如何反映在企业社会绩效上。基于国有企业（SOE）高管的双重职业轨迹视角构建理论框架，我们认为经济因素（绩效下降和薪酬差距）和政治因素（社会主义印记和政治职业视野）可能会不同地重塑激励方案对企业社会绩效的解释。利用“限薪令”作为准自然实验和对照实验揭示了对高层管理人员的薪酬限制会导致企业社会绩效的下降。当焦点公司继承了更强烈的社会主义印记，且高管拥有更长的政治前景时，这种关系会减弱。相反，当公司面临严重的绩效下降，且高管的薪酬相对于同行较低时，这种关系会加强。研究结果表明，促使国有企业能够参与社会领域，薪酬是与政治和经济因素结合的不可或缺的激励因素。我们讨论高层管理人员激励与激励安排的启示，以及政府目标如何影响高层管理人员对薪酬激励的反应，这些反应对长期社会贡献至关重要。

Introduction

Few topics in strategy and corporate governance remain as tangled as state-owned enterprises (SOEs) and corporate social responsibility (CSR). In this vein, CSR addresses a wide range of obligations a business has to society (Campbell, 2007; Carroll, 1979; Wang, Tong, Takeuchi, & George, 2016), and SOEs serve as an extension of the government as the mainspring of corporate social performance (Bai, Lu, & Tao, 2006; Lin, Cai, & Li, 1998). In this regard, prior scholars have paid growing attention to various forms of state role in corporate social performance, focusing on whether and why government control may promote CSR performance (Detomasi, 2008; Marquis & Qian, 2014; Wang & Qian, 2011). Relatedly, the government announced a series of guidelines and initiatives, emphasizing the importance of companies adjusting operational strategies to coincide with a harmonious society (Hofman, Moon, & Wu, 2017; See, 2009).

Two distinct perspectives offer insights into why SOEs outperform in the social responsibility area. A disproportion of research infers from the political view and assumes that SOEs will comply with political pressure, and the government plays a decisive role in propelling corporate involvement and performance in social issues (Li & Lu, 2020; Luo, Wang, & Zhang, 2017). In contrast, the agency view posits that SOEs maximize what dominant shareholders (the government) want at the expense of minor shareholders (Hsu, Liang, & Matos, 2023; Hu & Xu, 2022; Tan & Tang, 2016). Nevertheless, there remains a dearth of clarity regarding which analytical framework – be it political, economic, or a synthesis of both – holds greater relevance in elucidating the motivations underpinning SOEs' engagement in corporate social performance.

To address this issue, we integrate the political and agency views by adopting the dual-career tracks (economic and political) framework to incorporate the role of managerial incentive. In general, managers engage in a reflective process to assess and interpret their interactions with stakeholders, which serves to direct the attention allocation among their various duties (Basu & Palazzo, 2008; Ethiraj & Levinthal, 2009). Specifically, managers in SOEs strive to generally meet multiple objectives such as social and financial performance (Bai & Xu, 2005; Holmstrom & Milgrom, 1991; Zhou, Gao, & Zhao, 2017), and normative agency theory proposes that compensation is a basic pivot aligning the shareholders' interests with agents' interests (Cao, Lemmon, Pan, Qian, & Tian, 2019; Hu & Xu, 2022). With respect to social performance, our inquiry is directed toward understanding how executives concurrently assess managerial incentives through the dual lenses of political and economic considerations.

In this study, we analyze our research question using ‘Pay Ceiling Order’ (hereafter, the Order ¹) as the empirical context, providing an ideal exogenous shock to test the multiple incentives of SOEs' executives. The direct effect of the policy relates to the economic side of managerial incentive, which provides room to examine the hybrid effect of economic and political influence on social investment. In practice, to reduce the income gap in society², the Chinese Central State issued a set of rules to compress executive pay in SOEs in 2014. As such, the Order immediately changes the incentive arrangements of SOEs, and self-reflection incurred by the Order would motivate managers to reinterpret managerial incentive and modify resource allocation, especially in the social area (Basu & Palazzo, 2008; Flammer, Hong, & Minor, 2019; Su, Jiang, & Tian, 2020). The Order emphasizes the emerging goal of social equality, deprioritizing the political influence on conventional social issues. Meanwhile, the Order compresses the compensation of top executives, disincentivizing the economic alignment between executive engagement and societal interest.

Leveraging the dual-career tracks framework, we posit that political in tandem with economic factors reshape the reaction of top executives to CSR engagement in response to managerial incentive change. From the perspective of a political view, socialist imprints, and political career prospects as two crucial political influence constituents which could rectify the interpretation and resource allocation process (Li & Lu, 2020; Wang & Luo, 2019). First, Chinese SOE executives are quasi-officially paid as business managers but governed by state officials (Lin, 2016, 2017); they are given political promotion incentives and economic compensation incentives to drive their efforts (Cao et al., 2019; Wang, Fu, Huang, & Wang, 2014). Second, SOEs with inherent socialist imprints and top executives are inclined to carry political imprints propagated by the Party (Marquis & Qiao, 2020; Wang, Du, & Marquis, 2019; Xu, Zhou, & Chen, 2022; Zhang, Ren, & Wu, 2023). In contrast, from the economic view, the business prospect and current economic reward reshape top executives' interpretation concerning economic calculation (Bosse & Phillips, 2016; Wade, O'Reilly III, & Pollock, 2006). In particular, performance declines impose disciplinary pressures on top executives (Kuusela, Keil, & Maula, 2017), enhancing the sensitivity of top executives' reactions to the disincentive effect. The Order will change the relative pay gap among top executive peers, which tempts the top executive to reduce further efforts (Coles, Li, & Wang, 2018). Thus, these political factors – executive political career concerns and socialist imprints – would weaken the negative effect, while the economic factors – performance declines and relative pay gap – would strengthen the negative effect.

We tested our hypotheses using China's publicly listed firms from 2010 to 2018, and a controlled experiment with 79 participants, and the results largely support our conjectures. Our study offers two main contributions. First, this study sheds light on the dual-career tracks of SOEs' top executives, while the typical executive incentive literature that emphasizes the economic incentive as a primary driver (Cao et al., 2019; Hu & Xu, 2022; Jain, Zaman, & Harjoto, 2024). Our article thus provides a new examination of the Chinese state sector, in which socialist imprints and political career horizons reshape executive responses to economic incentives. Second, our study extends the role of SOE in CSR research by investigating the managerial interpretation of stakeholder relationships that vary on compensation arrangement. Existing research usually assumes that taking social responsibility is an inherent role of SOEs, especially when external demands increase (Jia & Zhang, 2013; Luo & Wang, 2021).

Theoretical Background and Hypotheses Development

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, 2020; Stan, Peng, & Bruton, 2014). 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, 2020; Raynard, Lu, & Jing, 2020). 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, 2015; Stan et al., 2014). 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, 2012; Li & Lu, 2020).

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, 2018; Su et al., 2020). 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., 2006; Tirole, 1994). 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, 2006a; Groves, Hong, McMillan, & Naughton, 1994). 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., 2019; Yang, Wang, & Nie, 2013; Zheng, Li, Xu, Lin, & Zhao, 2012). 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, 2011; Holmstrom & Milgrom, 1991; Jensen & Meckling, 1976). 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, 2016; Zhu, De Cieri, & Dowling, 1998). 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, 2015; Lin et al., 1998). 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., 2019; Groves et al., 1994).

Correspondingly, executive compensation has changed significantly along with the reform of the state sector in recent decades (Chan & Ma, 2011; Lin, 2016). 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, 2005). 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, 2022; Wang & Luo, 2019).

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, 2011; Jiang & Kim, 2020). 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., 2012; Mengistae & Xu, 2004). 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, 2011; Lin, 2016).

The multitasking nature of executive compensation of SOEs is well established, and many scholars investigate whether executive compensation matters (Bai et al., 2006; Holmstrom & Milgrom, 1991). 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., 2019; Conyon & He, 2011; Hu & Xu, 2022; Mengistae & Xu, 2004). 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., 2020; Hu & Xu, 2022). 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., 2020; Sun & Tong, 2003), and top executives in SOEs are quasi-officials in a political system and a labor market for professional managers (Hu & Xu, 2022; Manion, 1985). Thus, Chinese SOEs have two parallel personnel systems: regular corporate management and Party-state systems (Walder, 1995; Wang, Yi, Zhang, & Peng, 2022). 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.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., 2014; Yang et al., 2013).

Rooted in agency theory, dual-career tracks – business and politics – motivate SOE executives to balance multiple task requirements (Bai et al., 2006; Bai & Xu, 2005; Cao et al., 2019). 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., 2020; Zhang, Liu, & Cai, 2015). Finally, the incentives in each track present a somewhat substituted relationship (Cao et al., 2019; Chen, Guan, & Ke, 2013). For example, Cao et al. (2019) 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, 1995). 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, 2020; Liu & Xiao, 2015). For example, environmental protection performance carries more weight in the evaluation criteria of officials than in the past.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.

Political influence mechanism

The political view emphasizes the influence of the government, underscoring the intrinsic societal objectives embedded within the ultimate distribution of economic resources (Hsu et al., 2023; Jia & Zhang, 2013; Wang & Qian, 2011). In response to natural disasters, for example, SOEs expand the government to meet social desirability (Hofman et al., 2017; Jia & Zhang, 2013). The underlying reason is that SOEs were born with multitasks, including social and economic goals, and are endowed with a degree of discretion over operation and management decisions (Bai & Xu, 2005; Holmstrom & Milgrom, 1991; Stan et al., 2014). Relatedly, top executives of SOEs are semi-state officials whose compensation was mainly determined by the civil service pay system, implying equality between executives and ordinary employees (Chizema et al., 2015; Lin et al., 1998). Although the corporate privatization reform has transformed SOEs into economic engines boosting national economic growth, the essential objectives of SOEs in mitigating social conflicts and enhancing national welfare remain unchanged (Ji, Huang, & Li, 2021; Stan et al., 2014; Xu et al., 2022).

At its core, the Order signifies a governmental prioritization of addressing inequality, thereby underscoring the emergence of a novel social concern that swiftly ascends the agenda of policy imperatives. In general, when there are multiple tasks, compensation ‘serves to direct the allocation of the agents' attention among their various duties’ (Holmstrom & Milgrom, 1991: 25). In this light, the Order will divert the attention from normative social issues such as employee treatment and environmental protection, and then the political pressure on these social issues may divert firms' attention toward deciphering the government's agenda. Furthermore, the new emphasis would reduce the governments' monitoring effectiveness in propelling the SOEs to fulfill the social desirability concerning social responsibility. For example, SOEs are positioned to shift their primary focus from the provision of social welfare to the pursuit of efficient production during the transition process, and the executives can redirect their efforts from social welfare responsibilities toward the maximization of economic gains (Bai et al., 2006; Tirole, 1994). Integrating the inherent emphasis on social issues and distraction induced by the Order, we posit that SOEs might respond neutrally in the subsequent CSR performance.

Economic incentive mechanism

From an economic perspective, the agency theory posits that executive compensation represents a transactional process, wherein compensation is exchanged for the fulfillment of managerial duties aligned with shareholder value maximization. Likewise, SOEs also use compensation as an incentive to align the interests of executives and major shareholders (the government), and compensation increases with the marginal return of executive action (Conyon & He, 2011; Mengistae & Xu, 2004). Along with the strong economic growth of the state sector, SOE top executive compensation has increased significantly over past decades. However, such change is accompanied by rising pay differentials between top executives and ordinary employees (Chizema et al., 2015). While compensation is acknowledged as a pivotal determinant in driving managerial efforts (Conyon & He, 2011; Jensen & Meckling, 1976), the phenomenon of excessive executive compensation has been identified as potentially detrimental, undermining employee motivation, social equity, and overall firm performance (Shaw, 2014; Wade et al., 2006). To mitigate the tendency of overpayment in the state sector, the Order issued in 2014 imposes a maximum of 600,000–800,000 RMB on annual executive pay along with a maximum of ten times the pay differential between executives and workers of the SOE.⁵

It is imperative to recognize that the SASAC appraises the performance of SOEs through a dual-criteria framework, encompassing both financial and political dimensions (Du et al., 2012; Hu & Xu, 2022; Wei, 2021). As previously indicated, SOE executives now assume the dual role of state officials and business managers, with their compensation contingent upon their success in achieving both the financial goals of all stakeholders and the political mandates set forth by the state. In other words, they are now driven by economic (i.e., compensation) and political incentives (Cao et al., 2019; Hu & Xu, 2022). In practice, financial performance is inherently linked to objective metrics that offer limited scope for manipulation by top executives (Du et al., 2012). Conversely, political performance encompasses a broad spectrum of activities, providing executives with the discretion to modulate their engagement and investment of effort within this domain (Lin, 2016; Yang et al., 2013). However, some anecdotal evidence supports that a bonus ceiling on executive compensation would disincentivize managerial efforts to improve financial performance in private firms (Murphy, 2013; Murphy & Jensen, 2018). The compensation rearrangement might distort the effort allocation of SOE executives, and we contend that SOEs are more inclined to demonstrate less compliance with the regulatory directives, as manifested in their social performance, rather than through the more readily quantifiable metrics of financial performance.

Based on the abovementioned arguments, we contend that the confluence of political influence and economic incentive mechanisms predict a negative relationship between compensation restriction and the corporate social performance of SOEs, which engender a diminished motivational drive among executives to adhere to the intrinsic societally desirable objectives of their organizations. In light of the recent regulatory directive, the Order prioritizes social equality as a pivotal objective; there is an inferred de-emphasis on social issues that have been previously highlighted. Considering the disincentive effect of compensation restriction and final performance evaluation criteria, we posit that a concomitant decline in the resource allocation dedicated to CSR initiatives. Consequently, executives may perceive this regulatory shift as indicative of a diminished emphasis on economic incentives and political impetus, potentially leading to a diluted commitment toward the comprehensive enactment of corporate social responsibilities. Thus, we propose the following baseline hypothesis:

Hypothesis 1 (H1): Compensation restriction will have a negative effect on SOEs' CSR performance.

Unpacking the Two Parallel Perspectives

Thus far, our theorizing is built on the assumption that SOE executives are motivated by monetary rewards following the normative agency theory prediction and political influence fueled by the evaluation and consideration of political promotion. The literature on executive compensation suggests that executives differ in motives (i.e., what they most want and fear), and these differences will diverge the ‘valence’ of monetary rewards in motivating their efforts (Finkelstein, Hambrick, & Cannella, 2009). In developing the model of person-pay interaction, Wowak and Hambrick (2010: 804) recognized that ‘the role that individual differences play in shaping, or moderating, the effects of pay arrangements on organizational outcomes’.

In particular, top executives of SOEs have two career tracks to pursue, and we build a set of moderators to capture the heterogeneous effects among SOEs. To this end, we consider two representatives of economic factors – performance declines (disciplinary pressures) and relative pay gap (violation of the fairness belief) – will increase the valence of monetary incentives and thus strengthen the negative effect of compensation restriction on SOEs' CSR performance; two political factors – individual political prospect (executive political career horizon) and collective political values (i.e., exposure to communist imprints) – will decrease the valence of monetary incentives and thus attenuate the above negative effect of compensation restriction on SOEs' CSR performance.

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, 2000; Firth, Fung, & Rui, 2006b; Jordan & Audia, 2012). 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, 1981; You, Shin, & Chung, 2023; Zhong, Ren, & Song, 2022). For corporate social performance, constant resource investment is necessary to maintain the status quo of corporate social performance (Wang & Choi, 2013). 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, 2020). 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, 2007; Popli & Raithatha, 2023). 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, 2023; Hambrick, Finkelstein, & Mooney, 2005). 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., 2017; Mannor, Wowak, Bartkus, & Gomez-Mejia, 2016). 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:

Hypothesis 2 (H2): Performance declines will moderate the relationship between compensation restriction and SOEs' CSR performance, in such a way that the negative effect will be stronger when the top executives experience performance declines.

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, 2016; Devers, Cannella Jr, Reilly, & Yoder, 2007). In a social context, almost everyone cares about fairness to some degree and is more sensitive to underpayment than overpayment (Adams, 1965; Fong, 2010). 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, 2014; Wade et al., 2006). 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., 2006), 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., 2018; Fong, 2010; Fong, Misangyi, & Tosi, 2010). 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., 2010). 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., 2018; Cowherd & Levine, 1992).

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., 2013; Lin, 2017). 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:

Hypothesis 3 (H3): The top executive's relative pay gap will moderate the relationship between compensation restriction and SOEs' CSR performance, in such a way that the negative effect will be stronger when the top executive experiences a large relative pay gap.

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., 2019; Hsu et al., 2023). Immediate monetary incentives and future career opportunities are the two main components of career concerns that motivate executives' efforts simultaneously (Cao et al., 2019; Holmström, 1999). 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, 2022; Zheng et al., 2012). Social and environmental issues have become a priority in the state's goal (Zuo, 2015). 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., 2012; Liu & Xiao, 2015; Zheng et al., 2012). 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, 2020; Wang & Luo, 2019). 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., 2019; Hu & Xu, 2022). 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, 2015). 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, 2020; Wang, Zhu, Chen, & Luo, 2021). 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:

Hypothesis 4 (H4): The top executive political career horizon will moderate the relationship between compensation restriction and SOEs' CSR performance, in such a way that the negative effect will be weaker when the top executive has a long career horizon.

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, 1965: 154), the persistent effect of imprinting could influence current decision-making (Marquis & Tilcsik, 2013; Stinchcombe, 1965). 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, 2000; Lin et al., 1998; Marquis & Qian, 2014). 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, 2016; Raynard et al., 2020; Wang & Luo, 2019). 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., 2020; Zhang, Ren, et al., 2023). 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, 1989; Kriauciunas & Kale, 2006; Wang & Luo, 2019; Zhang, Ren, et al., 2023).

During the socialist training and socialization processes, which historically emphasized people-oriented and egalitarian preferences (Haack & Sieweke, 2018; Han & Zheng, 2016), 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, 2018; Zhang, Ren, et al., 2023), ‘training programs and wage systems are more internal and people-oriented’ (Kriauciunas & Kale, 2006: 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., 2020). 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, 2018; Kriauciunas & Kale, 2006). 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:

Hypothesis 5 (H5): Firm exposure to communist imprints will moderate the relationship between compensation restriction and SOEs' CSR performance, in such a way that the negative effect will be weaker when the firms are exposed to stronger communist imprints.

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

Figure 2. Conceptual model

Study 1: Archival Study for the Observed Effect

Utilizing an archival study is to paint a preliminary portrait of the declining corporate social performance in response to compensation restriction in SOEs. It is posited that as executives tend to be disincentivized toward the compensation restriction of the government's policy design, they possess a low engagement when processing such information since the compensation restriction is pertinent to interpreting a disincentive design. As such, pairing a disincentive motivation and government direction should result in a negative response to the subsequent social performance. At the same time, considering the executives occupy dual-career tracks, the political and economic tracks might reshape the interpretation of compensation restriction. Given that the political track aligns the purpose of executives with the government, while the economic track partially diverges the interests of executives with the government. Thus, if there are inherent political forces toward the political track, decreasing the executives' compensation will result in less negative evaluations of the compensation design as a function of the subsequent social performance. In contrast, if there are inherent economic forces toward the economic track, decreasing the executives' compensation will result in more negative evaluations of the compensation design as a function of subsequent social performance.

Archival Methods

Empirical context

We adopt a difference-in-differences (DID) design to examine the effect of compensation restriction on CSR performance by using the ‘Pay Ceiling Order’ as a quasi-natural experiment. The Order issued in 2014 provides a unique setting to test our model for three reasons. First, the Order is mandatory. It was promulgated by the Central Politburo of the Communist Party⁶ – the highest state agency in China. The aim is to mitigate the tendency of extremely high executive compensation in a group of Central SOEs rather than regulating CSR practices in SOEs. Second, the Order is exogeneous. It signals to top executives that their future compensation has a strict ceiling, which is beyond their negotiation capacity. And it does not impose additional changes on the evaluation criteria and process of SOE executives. Third, the Order was carried out in a group of Central SOEs (the treatment group), allowing us to use the rest of the SOEs and private firms as the control group of Central SOEs. Notably, the Order recommended the local SOEs should also be actively and steadily promoted with reference to the spirit of the compensation reform.⁷ As such, we also test the effect on all SOEs (the treatment group) with private firms as the control group. Therefore, the ‘Pay Ceiling Order’ provides a unique setting to investigate whether and how top executives respond to compensation arrangement by changing the resource allocation in social performance.

As with the corporate management system everywhere, top executives in listed SOEs commonly include the CEO, Vice-CEO, CFO, the board of directors, a chairman, and independent board members. While all these executives matter, the Chairman of the Board of Directors (CBD, Yi-ba-shou) and the CEO (Er-ba-shou) are considered as the most influential top executives in Chinese firms (Cao et al., 2019; Yang et al., 2013). We thus focus on the Chairman of the board and the CEO as the representatives of the top executives who are subjected to compensation restrictions in our main analysis.

Sample and data

Our sample consists of all SOEs and private-owned enterprises listed on the Shanghai and Shenzhen stock exchanges between 2010 and 2018, holding four years surrounding the enforcement of the Order to observe the reactions. Information on firms' CSR was drawn from the HeXun (www.hexun.com), the first vertical financial portal and one of the largest financial and securities information service providers in China. HeXun provides CSR scores based on firms' annual CSR reports and annual reports. Other CSR ratings, such as the Rankins CSR ratings (RKS), cover only firms that issue annual CSR reports. The HeXun CSR index allows us to include firms that did not publish CSR reports during our research window. It also has a theoretical stakeholder framework and adopts a unified rating scheme to calculate the total CSR score. With a broader range of information sources and more comprehensive CSR ratings, the HeXun CSR index has been widely used in previous studies (e.g., Cheng, Chu, Deng, & Huang, 2022; Li & Guo, 2022). We thus use the HeXun CSR index in our main analysis. In robustness checks, we use CSR data offered by the Information Technology Co., Ltd (GTA), another widely used source of CSR information in previous research (Zhang, Wang, & Zhou, 2020; Zhang, Zhou, & Lyles, 2023).

Information on firm-level and executive-level characteristics was collected from the Chinese Stock and Market Accounting Research (CSMAR) database. We exclude financial firms because their disclosure requirements and accounting rules are significantly different from those of firms operating in manufacturing industries. Firms that are specially treated (ST or *ST) or delisted during our research window are excluded. After excluding observations with missing information, our final sample consists of 21,636 firm-year observations (3,500 firms).

Measures

Dependent variable

CSR performance. The Hexun dataset provides CSR performance on five major dimensions: investors, employees, suppliers, customers, and environmental and social responsibility. Following previous studies (Li, Xu, McIver, Liu, & Pan, 2022; Xu & Ma, 2022), we measure firm CSR performance as the sum of these five dimensions.

Independent variables

Treat. Using the 2014 Order as the exogeneous event, we first identify a group of 357 Central SOEs whose top executives are subjected to the Order as our treatment group, and we use the rest listed firms as the control group. Treat is set to 1 if a firm was identified as the treatment group; otherwise, it is set to zero. Because the Order also spreads to local SOEs, as the suggested announcement, to improve the validity of the hypotheses, we conduct additional analyses where we identify Central SOEs and local SOEs (totaling 1,029 firms) as the treatment group while using the rest of the private firms as the control group.

Post. We take 2014 as the event year, as the Order was formally enforced on January 1, 2015. Then, we compare the treatment and control groups in the pre-event period (2010–2013) versus the post-event period (2015–2018). Thus, Post is set to 1 for fiscal years 2015 to 2018 and zero for 2010 to 2014.

Moderators

Performance declines

Consistent with previous studies (Audia et al., 2000; Kuusela et al., 2017; You et al., 2023), we measure the performance decline variable in three steps. First, we compute industry-adjusted return on assets (ROA⁸) performance to capture operating performance. Second, we compute a difference measure by subtracting the current year's adjusted performance from the previous year's adjusted performance. Third, we invert the coding of the difference in order to make the results easier to understand. The larger value indicates that the firms are experiencing a severe performance decline.

Executive's relative pay gap

Given that every top executive (i.e., CEO, Chairman) in the same province⁹ has the incentive to compete for the executive position with the highest pay, they are familiar with each other in terms of regional interaction (Coles et al., 2018; Jain et al., 2024). We calculate the compensation gap between the highest total executive pay (cash pay) and the executive under consideration in the same province based on the two-digit Chinese National Industry Classification (CNIC) codes and the focal executive's total pay. Then, we measure the relative pay gap using the natural logarithm of the relative distance plus one (Coles et al., 2018; Jain et al., 2024).

Executive political career horizon

In line with previous studies (Cao et al., 2019; Li & Lu, 2020; Wang & Luo, 2019), age is a definitive criterion for a potential candidate to retire (at age 68). The top executives of Central SOEs have similar political status, and we adopt a conservative measure to capture the political career horizon as the number of years before the CEO or Chair retires. This reflects the amount of time remaining until the executive reaches retirement age (Matta & Beamish, 2008): the greater the number, the higher the possibility of a manager's political promotion.

Executive exposure to socialist imprint

Following previous studies about socialist imprint (Han & Zheng, 2016; Marquis & Qiao, 2020; Wang et al., 2019; Xu et al., 2022; Zhang, Ren, et al., 2023), we contend that pro-market institutions could significantly change firms' decision making and create socialist imprint variable through a continuous variable as the number of years between 1992 and the year a firm had founded. To easily interpret, we then reverse-coded the continuous number so that the greater the value of this variable, the stronger the socialist imprint.

Control variables

To rule out alternative explanations, we controlled for a vector of characteristics that might influence the firms' social performance (Fu, Tang, & Chen, 2020; Luo & Wang, 2021; Marquis & Qian, 2014; Zhang, Marquis, & Qiao, 2016). We control for five typical firm-level variables that have been suggested in previous studies to influence our dependent variables. Firm size is measured as the natural log of total assets (McWilliams & Siegel, 2001; Zhang, Chen, & Jia, 2023). Firm age (the logarithm of number of years since the firm was founded) is included to control for the effect of firm life-cycle (Cuypers, Koh, & Wang, 2016). Advertising intensity (the ratio of selling expenditures to sales) is included to control for firm visibility (Wan, Wang, Geng, & Huang, 2023; Wang & Qian, 2011). Two variables, ROE and Tobin's Q, are included to control for firm financial performance (Li & Lu, 2020).

We control for several characteristics of corporate governance (Aguinis & Glavas, 2012; Chen, Crossland, & Huang, 2020). Two variables, share concentration (the proportion of shares owned by the top 10 shareholders) and percentage of independent board directors, are included to control for the monitoring effect of corporate governance on CSR (Lau, Lu, & Liang, 2016; Li & Lu, 2020). Managerial ownership, measured as the natural log of 1 plus the total shares owned by a firm's top executives, is included to control for the potential incentive effect of corporate governance on CSR (Flammer et al., 2019). Leverage (the natural log of 1 plus the total debt scaled by total assets) is included to control for external monitor that may tempt firms to change the degree of CSR engagement (Lin, Tan, Zhao, & Karim, 2015). We also include big-4 to control for market intermediaries that may influence the transparency and quality of firm CSR (Li, Li, & Wang, 2019). Big-4 is a dummy variable that equals 1 if a firm's audit firm is any of PwC, Deloitte, EY, or KPMG and zero otherwise.

In addition, a dummy variable, mandatory disclosure, is included to control for institutional influences on CSR quality (Chen, Hung, & Wang, 2018; Zhang et al., 2023); It equals 1 if a firm is mandated to issue a CSR report and zero otherwise. Finally, we include both firm- and year-fixed effects to control for any potential unobserved time-invariant factors across different years of observations.

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.

Results

Table 1 displays the descriptive statistics and correlation among variables in the subsequent empirical analyses. The correlations between CSR performance and SOEs are positive and significant, which provides preliminary support for the social obligation arguments. The average portion of central and all SOEs is 13% and 41%, respectively.

Table 1. Descriptive statistics and correlations analysis

Notes: Using list-wise deletion to handle missing values. Correlation coefficients greater than 0.01 are significant at p < 0.05 level.

Table 2 reports the empirical analyses for compensation restriction and CSR. Given the Order will spread to all SOEs rather than only targeting the Central SOEs, we analyze and compare the effects on these two samples separately in the following regressions. As shown in Table 2, the coefficient of Treat × Post in model 2 is negative and significant (β = −4.920, p < 0.01), indicating that compensation restriction leads to significant CSR performance declines among Central SOEs. The coefficient of Treat × Post in model 4 is also negative and significant (β = −3.345, p < 0.01), suggesting that compensation restriction leads to significant CSR performance declines among all SOEs (i.e., Central and local SOEs). Thus, H1 is supported.

Table 2. Compensation restriction and 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.

We next explore the boundary conditions of the relationship between compensation restriction and CSR. The main purpose is as follows: While the mechanism proposed in the arguments might be implausible to measure directly, we use the moderators to validate it somewhat. The idea of the dual-career track suggests that the top executives of SOEs are economic and political humans (Hu & Xu, 2022; Zhou et al., 2017), providing a comprehensive framework to incorporate both political and economic factors to identify the motivation change surrounding the compensation restriction.

H2 predicts that performance declines exacerbate the negative effect of compensation restriction on CSR performance. As shown in Table 3, the coefficient of Treat × Post × Performance declines in model 2 is not significant (β = 3.317, p > 0.1); the reason for the result might be that the Central SOEs are not sensitive to financial performance change. By contrast, the coefficient of Treat × Post × Performance declines in model 4 is negative and significant (β = −21.204, p < 0.05), indicating that firms facing performance decline situations exacerbate the negative effect of compensation restriction, especially for all SOEs. We speculate the reasons behind the differences come from the resource endowment and evaluation criteria (Cao et al., 2019; Wei, 2021). Therefore, H2 is not fully supported.

Table 3. Moderating effect of performance declines

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.

H3 predicts that the executive's relative pay gap exacerbates the negative effect of compensation restriction on CSR performance. For the sake of robustness, we separately present the results for CEO and chairman of the board of directors. For CEO, as shown in Table 4, the coefficient of the Treat × Post × Pay gap in model 2 is negative but not significant (β = −0.971, p < 0.05), indicating that the CEO's relative pay gap does not exacerbate the negative effect of the compensation restriction in Central SOEs. Meanwhile, the coefficient of Treat × Post × Pay gap in model 6 is negative and insignificant (β = −0.051, p > 0.1), suggesting that the CEO's relative pay gap does not consistently exacerbate the negative effect of compensation restriction in all SOEs. For chairman of the board of directors, analyses of the Chairman's relative pay gap reveal a different pattern. In Table 4, the coefficient of Treat × Post × Pay gap in model 4 is β = −0.789, p < 0.1). This suggests that the Chairman of the board's relative pay gap exacerbates the negative effect of compensation restriction on CSR performance in Central SOEs, proving some support for H3. Yet, the coefficient of the Treat × Post × Pay gap in model 8 is insignificant (β = −0.103, p > 0.1); this suggests that the Chairman of the board's relative pay gap does not consistently exacerbate the negative effect of compensation restriction in all SOEs, providing no support for H3. Overall, the executives of Central SOEs are more sensitive to the relative pay gap than all SOEs, which might be the fairness is imperative for individuals in high-ranking status. Thus, H3 is not fully supported.

Table 4. Moderating effect of the relative pay gap

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.

H4 predicts that the top executive political career horizon alleviates the negative effect of compensation restriction on CSR performance. As shown in Table 5, the coefficient of Treat × Post × Political career in model 2 is positive and significant (β = 0.528, p < 0.01), indicating that CEO career horizon weakens the negative effect of compensation restriction in Central SOEs. Also, the coefficient of Treat × Post × Political career in model 5 is positive and significant (β = 0.467, p < 0.01), suggesting that CEO career horizon weakens the negative effect of compensation restriction in all SOEs. These results provide support for H4. Analyses of the chairman's career horizon reveal similar findings supporting H4. In Table 5, the coefficient of Treat × Post × Political career in model 4 is 0.372 (p < 0.1), and this is 0.289 (p < 0.01) in model 8, suggesting that the Chairman of the board's career horizon weakens our main effect. Thus, these results provide strong support for H4.

Table 5. Moderating effect of political career horizon

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.

H5 predicts that executive exposure to socialist imprint alleviates the negative effect of compensation restriction on CSR performance. Table 6 reports regression results for the moderating effect of executive exposure to socialist imprint. Given our conservative measure of socialist imprint proxied by communist imprints of the firm where an executive works, the value of this variable for the CEO and the Chairman of the board of the same firm will be the same. We thus do not conduct separate analyses. As shown in Table 6, the coefficient of Treat × Post × Socialist imprint in model 2 is positive and significant (β = 0.433, p < 0.05), indicating that executive exposure to communist imprints weakens the negative effect of compensation restriction in Central SOEs. Likewise, the coefficient of Treat × Post × Socialist imprint in model 4 is positive and significant (β = 0.350, p < 0.01), suggesting that executive exposure to communist imprints weakens the negative effect of compensation restriction in all SOEs. Thus, these results provide support for H5.

Table 6. Results for moderating effect of the socialist imprint effect

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.

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, 2022), 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, 2022; Chatterji, Durand, Levine, & Touboul, 2016). 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., 2020; Zhang, Zhou, et al., 2023) 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, 2022), 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, 2017; Hawn & Ioannou, 2016; Hsu et al., 2023). 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., 2012; Zuo, 2015). 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., 2006; Lin et al., 1998). 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., 2019). 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.

Employees' compensation change

In addition to the effect on top executive compensation, we examine the effect on employee compensation (Kong, Kong, & Lu, 2020; Wang, Dai, & Kong, 2021). Given the top executive control and determine the final allocation of subordinate compensation, the compensation restriction policy might trigger the social comparison motivation. In this regard, the top executives could reduce employee compensation simultaneously. If that is the case, we will see a significant reduction in employee compensation after the enforcement of the Order. We re-run the regression in terms of equation 1, and the results show that the employee compensation does not reduce as predicted. We argue that SOEs realize the underlying purposes of the Order and comply with the common prosperity calling.

Based on the above results, we could infer that restricting executive compensation could ameliorate within-firm income inequality (Dittmann, Maug, & Zhang, 2011), consistent with the purported effect. Unfortunately, the compensation of ordinary employees does not significantly increase at the same time. Accordingly, the compensation restriction order could fulfill the common prosperity goal to some extent because the within-firm income inequality accounts for more than one-third of the income inequality of the whole society (Song, Price, Guvenen, Bloom, & Von Wachter, 2019).

Conclusion

The findings of Study 1 support the speculation according to dual-career tracks, in which the pollical factors attenuate the negative relationship between compensation restriction and corporate social performance while the economic factors accentuate it. In sum, the executives of SOEs are both business managers and politicians. The compensation restriction could only achieve one goal of reducing social inequality while losing another social good as a cost.

However, a key limitation of our archival study is that it was unable to capture the proposed underlying mechanisms associated with executives' motivational and attitudinal response to compensation rearrangement, advice giving and creativity. To address this gap, we conducted a controlled experiment study whereby we aimed to replicate our main findings, establish causality, and further test the underlying mechanism that emerge from executives' motivation and perception.

Study 2: Experimental Study for Mechanisms and Causation

An archival study cannot conclusively show causality and directly tackle mechanisms. Therefore, we designed an experimental study to establish a causal relationship between compensation restriction and corporate social performance. One aim is to replicate the main finding of the above archival study: compensation restriction incurs effort distortion in regard to work engagement and reinterpretation in regard to social pressure associated with political consideration. Apart from the replication, we also aim to provide evidence of the causal mechanisms of these disincentive interpretations.

Basic Procedure

We conducted an experimental vignette consistent with prior studies, as shown in the Appendix (Alesina, Ferroni, & Stantcheva, 2021). To balance the internal and external validity, the participants were MBA students (Audia et al., 2000; Krause, Whitler, & Semadeni, 2014). 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.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, 2018; Ong, Mayer, Tost, & Wellman, 2018). Accordingly, we use Ong et al. (2018) 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, 2023). 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, 2011; Lin, 2010; Schaufeli, Bakker, & Salanova, 2006). In this study, we use Lin's (2010) 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., 2023; Scott & Barnes, 2011). In this study, we use Kong et al.'s (2023) 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, 2003; Yang, Wang, Zhou, & Jiang, 2019). In this study, we use Yang et al.'s (2019) 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, 2016; Exley, 2018). 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, 2018; Flammer & Kacperczyk, 2019; Krause et al., 2014). 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, 2024). 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.

Discussion

This study explores and finds a negative effect of compensation restriction for top executives of SOEs on CSR performance. Historically, SOEs are a kind of agent to take responsibility for social issues, and the dual-career tracks of top executives determine the managerial interpretation of compensation arrangements. According to the political view, socialist imprints and political career concerns rectify the disincentive effect and direct the firms' resources to enhance social performance, and a compensation restriction will not immediately disincentivize them in social performance. According to economic view, however, top executives in SOEs also have economic motivation in the business track, and disciplinary pressure induced by performance declines and unfairness perception triggered the low pay ranking in peer executives, further prompting a disincentive effect. This study uses the Order as a quasi-experimental setting and a controlled experiment to empirically test the economic incentive and joint effect with political and economic factors in CSR performance. Additionally, we find a significant effect in improving income equality within a firm, which could provide preliminary suggestions for future policy enactment and enforcement. In sum, the policy might catch one goal (i.e., societal equality) while losing another goal (i.e., social performance).

Theoretical Contributions

This study contributes to the literature on CSR performance and SOE incentives. First, we contribute to CSR performance literature by exploring the compensation restriction policy from an individual perspective. Prior research investigates institutional pressure and shareholder conflicts in determining the CSR performance of SOEs (Hu & Xu, 2022; Li & Lu, 2020; Marquis & Qian, 2014); another stream of research focuses on the compensation arrangement including CSR requirements (Flammer et al., 2019; Maas, 2018). However, little research pays attention to the compensation regulation for SOEs and the boundary role of political and economic factors related to dual-career tracks. To fill this gap, we explore the compensation restriction policy issued by the central government, which provide a quasi-experimental context to qualify the effect of economic incentive on SOEs' top executives. The empirical findings support that the economic disincentive effect will negatively affect CSR performance, even though the SOEs have an inherent purpose of social obligations. In doing so, this study complements extant research emphasizing political influence in SOEs' CSR engagement.

Second, we also broaden the understanding of agency theory by directly incorporating political and economic factors of dual-career tracks into conceptual and empirical models. The separation of ownership and control rights leads to a series of agency issues, and prior research of SOEs has shown a government intervention and private–government principal conflicts in finance performance (Bai & Xu, 2005; Hu & Xu, 2022; Zhou et al., 2017). However, we need a more comprehensive understanding of the dual-career tracks in social performance. Based on the compensation restriction policy, we emphasize the factors in political career tracks that could complement the economic prediction of agency theory.

Third, we enrich the SOE incentives literature by theorizing a novel way in which managers interpret the change in compensation arrangement. Prior research on SOE incentives focuses on pay-for-performance sensitivity and perks (Firth et al., 2006a; Groves et al., 1994; Mengistae & Xu, 2004). However, the imprinted values and relative pay gap receive little attention in this stream of literature. Based on typical multitasking agency theory (Ethiraj & Levinthal, 2009), we posit that top executives make sense of social and economic goals by weighting either political or economic concerns, depending on the context. The economic incentive has an influential effect on the interest alignment of shareholders (i.e., the state) and managers (Hu & Xu, 2022), and promotion-based incentives (i.e., political career concerns) and socialist imprints also show a decisive effect (Wang & Luo, 2019; Xu et al., 2022).

Practical Implications

In addition to theoretical contributions, this study provides practical implications for policymaking and corporate governance. First, the policymaking process should consider both short-term incentives and long-term performance when arranging top executives' compensation. Given the nature of dual-career tracks for SOE executives, economic incentives account for contemporary work engagement, while political incentives might benefit the broader society, such as social welfare. Given that the economic incentive is a kind of explicit incentive strategy and the promotion-based incentive is a kind of implicit incentive strategy, the organizational attributes and prospective career concerns are indispensable in the design of the incentive arrangement. In this respect, SOE executives are both economic and political men, considering the economic and political tracks into interpretation of resource allocation process. Although the SOEs are responsible for social issues by default, the economic incentives influence the dedication to social obligations as well. In this light, policymakers should value policy according to incentive compatibility in compensation arrangement.

Second, corporate governance and executives should pay attention to the balance between short-term and long-term goals, and the internal and external governance systems ensure the work engagement of top executives. Job security and managerial reaction are crucial factors for top executives, especially when the compensation arrangement changes by restricting their compensation. In practice, achieving the goal of common prosperity in the whole society should not purely restrict the income of the high-income group; the effect of ‘splitting the pie’ is not usually effective. Instead, more policies related to ‘growing the pie’ for both the high-income and low-income groups and providing the latter with a higher income growth rate might be plausible to achieve the goal.

Limitations and Future Research Directions

Our study also has several limitations that provide opportunities for future research. First, the study uses compensation restriction policy as an exogenous shock to explore the firms' reaction; the economic and political forces behind the reaction need more nuanced settings to capture them fully. Future work could examine the process of managerial interpretation if there are available qualitative archives involving the reaction of top executives. Second, we still cannot exclude the possibility of information leakage before the formal announcement of the policy, although we employ the DID approach to examine the policy effect. We designed a controlled experimental study and obtained qualitatively similar results, and we hope future work can mitigate the alternative channel of inside information to deepen the understanding. Likewise, the empirical results show an inconsistent pattern of SOE reactions, the future research needs more details to disentangle the nuanced differences between Central SOEs and local SOEs in response to governmental policy. Finally, we claim that top executives have two kinds of motivations in the regular decision-making processes, a conceptual problem arising from not being able to observe the hiring process at a more granular level. Future research could use interviews or survey methods to dive into the underlying motivations and directly measure the cognitive and motivational reactions during resource allocation processes.

Conclusion

Our findings show that compensation is an important force that drives top executives in SOEs and affects resource allocation among various social issues. Meanwhile, two components of political concerns – socialist imprint and political career horizon – significantly attenuate the disincentive effect of compensation restriction on CSR; two components of economic concerns – performance declines and relative pay gap partially accentuate the disincentive effect of compensation restriction on CSR. Our investigation delves into the integrative motivations underpinning SOE social responsibility behaviors, encompassing both economic and political views. By doing so, our research offers a holistic framework that elucidates the nuanced heterogeneity observed in SOEs' approaches to social responsibility.