The Origin of Proactive Environmental Corporate Social Responsibility (ECSR) of Large Firms: Institutional Embeddedness—Driven, Family Involvement-Promoted, or Resource-Dependent?

Why are some firms in China highly engaged in proactive environmental corporate social responsibility (ECSR) while others are not? Knowledge of the determinants of proactive ECSR is a key research question among academics and practitioners. This study seeks to advance the ECSR literature by suggesting a configuration way of integrating institutional embeddedness, family involvement, and the resource base factors of firms to predict proactive ECSR investment, a proxy for proactive ECSR. Specifically, this study explores how the combination of institutional embeddedness, family involvement, and the resource base factors of firms jointly shape the proactive ECSR investment of large firms. Using fuzzy-set qualitative comparative analysis (fsQCA) on a sample of 162 large private firms in China, this study finds that no condition alone is sufficient to account for high proactive ECSR investment. Institutional embeddedness, family involvement, and the resource base factors of firms must interact with one another to produce highly proactive ECSR investments. There exist two distinctive configurations that provide a nuanced picture of the drivers of proactive ECSR investment. These findings provide meaningful insights for theory and practice and pave the way for future research in the domain of ECSR.


Introduction
The role of business in society has changed profoundly in the past few decades [1]. Firms have long been regarded as having potential negative impacts on the environment due to the damage of excessive industrial production to the ecological environment [2]; currently, more and more firms are undertaking proactive environmental corporate social responsibility (ECSR) (business practices adopted voluntarily by firms that go beyond regulatory requirements to support environmental sustainability) [3]. Nowadays, proactive ECSR is increasingly becoming a worldwide phenomenon.
However, we must identify the reason why some firms take more proactive environmental stances than others to encourage more firms to engage in proactive ECSR. Not surprisingly, there have been numerous studies examining factors affecting ECSR. Present studies, nonetheless, have focused on either the role of institutional embeddedness (e.g., [4][5][6][7]) or the effect of family involvement (e.g., [2,[8][9][10]) (taking the internal resource base factors of firms as control variables), and little attention has been devoted to examine the total and configuration impact of institutional embeddedness and family involvement on firms' adoption of ECSR [11]. Moreover, in terms of institutional embeddedness, existing studies have mainly focused on the effects of formal institutional embeddedness and individual-level institutional embeddedness (e.g., political membership of entrepreneurs; see [4,7,12,13]) on ECSR, largely ignoring the impacts of informal institutional embeddedness (e.g., industrial association membership) and organization-level Political membership embeddedness is a form of formal and individual-level institutional embeddedness, party branch embeddedness is a form of formal and organizational-level institutional embeddedness, and associational membership embeddedness is a type of informal and individual-level institutional embeddedness. Furthermore, political membership embeddedness and party branch embeddedness are distinct forms of political embeddedness, while associational membership embeddedness can be viewed as semipolitical embeddedness [15].

Proactive ECSR
In reality, companies exhibit different levels of commitment to ECSR. Following existing studies [3,17,29], ECSR can be divided into proactive ECSR (voluntary ECSR or non-obligatory ECSR) and passive ECSR (reactive ECSR or obligatory ECSR). The former refers to business practices adopted voluntarily by firms that go beyond regulatory requirements to support environmental sustainability, while the latter describes business practices adopted by enterprises to achieve compliance with government-imposed environmental regulatory standards with minimum efforts. Examples of proactive ECSR practices include establishing environmental performance goals, publicly disclosing environmental performance information, making voluntary environmental investments, and performing internal and external environmental audits [29], while typical example of passive ECSR practices is paying environmental pollution control fees.
Despite increasing voices urging businesses to fulfill their obligations to the environment, many firms are unwilling to engage in proactive ECSR due to several barriers (e.g., lack of incentives or resources, costly with an uncertain payoff) [30]. However, there are still many firms undertaking proactive ECSR. Hence, the underlying reasons (driving forces) for firms' engagement in proactive ECSR are well worth exploring. As such, this paper focused on proactive ECSR. Following Zhou et al. [17], the proactive ECSR was measured by the voluntary environmental investment of firms for pollution abatement, such as sewage treatment, air pollution treatment, soil pollution treatment, solid waste treatment, flue gas treatment, and waste recycling.

Institutional Embeddedness and Proactive ECSR
The drivers of ECSR have been a key research theme in the ECSR domain. Early researchers have attributed CSR (including ECSR) practices to altruistic motives and citizenship behavior (see [31][32][33]), while some recent literature (see [34][35][36]) has advanced the view that CSR practices could be driven by instrumental and strategic motives [16]. The strategic motive perspective is supported by the claims of legitimacy theory (LT). According to LT, firms continually strive to ensure that they are functioning within the bond and norms of the society in which they operate because the society grants firms without an inherent operation right a "license to operate" [37,38]. Hence, firms are strongly motivated to maintain their legitimacy through various practices, including proactive ECSR engagement. That is, the engagement of firms in proactive ECSR can be viewed as a response to institutional pressures [16].
The sources of the institutional pressures of firms are diversified, while scholars generally believe that pressures from political membership embeddedness, associational membership embeddedness, and party branch embeddedness are the main sources of the institutional pressures faced by Chinese firms [6,[14][15][16][17]27]. In modern China, various institutional embeddedness has become a common phenomenon, and the reasons are twofold. First, although China has made great progress in the transition toward a more market-oriented economy over the past few decades, its economic system still has a strong feature of government control [6]. Recent development in the literature has supported the notion that the greater the influence the government has on the operation environment of a firm, the more likely the firm is to engage in political strategies (e.g., [39][40][41]). Thus, it is easy to understand why it is fairly common to see extensive political connections between governments at all levels and firms [6]. That is, politically embedded firms seek to enjoy preferential treatment over government loans, contracts, licenses, and other scare resources controlled by local bureaucrats [6,27]. Second, to achieve political stability and socioeconomic development, on the one hand, the state strives to recruit business elites into the political system and industry associations organized by the government (known as "cooptation" or "absorption;" see [42][43][44][45][46]); on the other hand, the ruling party (the Communist Party of China, CPC) continuously embeds party branches into private enterprises. In this process, private entrepreneurs enter the political field and further seek protection and stability.
We contend that political membership embedded firms (serving as members of political councils such as the National People's Congress (NPC) or the Chinese People's Political Consultative Conference (CPPCC); see [27,47]) are more likely to engage in ECSR due to three reasons. First, one of the main characteristics of the environmental governance of China is the various social mobilizations taken by the government to make relevant social groups cooperate with government actions, while political membership embedded firms are usually the first object to be mobilized [48,49]. On the one hand, the government often first turns to political membership embedded firms because political mobilization usually follows the principles of "near before far," and political connection builds a bridge between the state and private firms and, thus, shortens the information and relationship distance between the two [7]. On the other hand, political membership embedded firms are more likely to acquiesce to demonstrate loyalty and secure benefits derived from close government-business linkages [16]. In this sense, political membership embeddedness provides a formal channel for the government to shift and decompose environmental governance tasks to various companies [7]. Therefore, political membership embedded firms are more likely to undertake environmental protection responsibility, and thus, actively engage in proactive ECSR practices.
Second, political membership embedded firms feel a stronger pressure to adopt ECSR practices which have been explicitly promoted by the Chinese government. Responding to government signals is a critical form to maintain political legitimacy [6,27]. The political identity of private entrepreneurs can increase the benefits brought by complying with government signals and correspondingly increase the costs (loss of legitimacy and other social and economic benefits) produced by non-compliance with government signals [15]. That is, political membership embedded firms are more sensitive to changes in the goals and priorities of the government [7]. Considerable facts indicate that the Chinese government has been signaling that ECSR is a desired activity over the last two decades [6,50]. For instance, the 11th Five-Year Plan for National Economic and Social Development in 2006 articulated a national vision rooted in the principles of developing a harmonious society and achieving sustainable development [51], which was widely seen as a signal of the transition of Chinese government policy from an economic growth at all costs to an economic growth balanced with tackling pressing social and environmental problems properly [27]. All the statements and guidelines of the government on CSR can be considered legitimacy guidelines (signals from powerful actors that define the appropriate activities of firms; see [52]) [27], and these imply that the institutional expectations from the Chinese government on firms' adopting CSR practices are particularly strong [16]. To maintain political legitimacy in the eyes of the government, political membership embedded firms are motivated to adhere to the government's signals and engage in proactive ECSR activities.
In addition, political membership embedded firms are more likely to receive social and government attention than their counterparts [13]. Specifically, political membership embedded firms are known by more people due to the high-profile political identity of their leaders. For instance, when the concurrent national meetings of the NPC and CPPCC (together known as "lianghui") are jointly held in China, the NPC deputies and CPPCC members are commonly the focus of media attention [53] Thus, these embedded firms are expected by the public to have high moral standards and do what is morally and legally respected [13]. Moreover, the government tends to take political membership embedded firms as pioneers and leaders in implementing the encouraged policy to play a guiding role for other firms [54]. The pressures and expectations from the public and the government may motivate these embedded firms to act more pro-environmentally, namely, practicing proactive ECSR activities. These arguments lead us to the following hypothesis: Hypothesis 1. Political membership embedded firms will make more proactive ECSR investments than non-embedded firms.
It has been a common phenomenon that the party organization of the CPC is embedded into private firms [55]. There are solid policy and legal bases for the establishment and construction of the party organization of the CPC. For instance, Article 30 of the Party Constitution of the CPC clearly stipulates, "Enterprises, rural areas, government organs, schools, research institutes, communities, social organizations, companies of the People's Liberation Army (PLA), and other basic units shall establish grassroots party organizations if there are more than three formal party members in these areas and/or institutions." As another example, Article 19 of the Company Law of China clearly states, "The party organizations of the CPC may be established in companies and carry out their activities in accordance with the Party Constitution of the CPC. Companies shall provide the party organizations of the CPC with conditions necessary for carrying out their activities." In fact, private firms have set up the party organizations of the CPC since the 1990s [45], and 1.88 million private firms have established the party organizations of the CPC by 2017 [55].
Researchers believe that the establishment of the CPC branches in private firms is not only a means of pursing legitimacy [14], but also provides the firms a channel to participate in politics and express their demands of interests [56]. We assert that party branch embedded firms are more likely to engage in ECSR due to an "environmental policy compliance effect" and "environmental values shaping effect." Specifically, private firms are less motivated to undertake environmental responsibility due to their property rights and profit-making purposes. Notwithstanding, as a formal institutional embeddedness, party branch embeddedness increases the pressure of firm legitimacy and, thus, prompts firms to follow the policies publicized by the CPC and the government. As mentioned earlier, the government has been promoting CSR in the last two decades. In fact, the CPC has also been promoting CSR over the same time. For instance, promoting "social responsibility among citizens, enterprises, and all kinds of organizations" has been documented by the statements of the Sixth Plenum of the 16th CPC Central Committee Congress [57]. To maintain political legitimacy, party branch embedded firms are more likely to respond to these signals and engage in proactive ECSR activities. In addition, party branch embeddedness can shape the environmental values of embedded firms through the long-term policy propaganda of the CPC and the active promotion of party members. Nowadays, the expectations of stakeholders associated with ECSR are continuously rising [58]. According to the Party Constitution of the CPC, the CPC is the representative of the fundamental interests of the overwhelming majority of the Chinese people, and all party members shall play exemplary and vanguard roles in various fields. Therefore, the embedding of the party branch of the CPC into private firms could effectively correct the deviation of their economic interestoriented principles (from economic interest-oriented to stakeholder interest-oriented), improve the environmental values of firms, and drive firms to engage in proactive ECSR, which leads us to formulate the following hypothesis: Hypothesis 2. Party branch embedded firms will make more proactive ECSR investments than non-embedded firms.
Industrial associations are generally regarded as an important form of the self-management of firms within an industry. In China, there are two major industrial associations: one is the All-China Federation of Industry and Commerce (ACFIC, also known as "Quanguo Gongshanglian"), and the other one comprises sector-based industrial associations (known as "hangyexiehui") [14]. The ACFIC was founded by the United Front Work Development (UFWD) of the CPC, and it works under the leadership of the CPC with the key functions of promoting the policies of the party among private industrialists and business owners [43]. Hence, as an officially sponsored business association, the ACFIC was born within the existing political system (tizhinei) [59]. Although membership in the ACFIC is voluntary, the members of the ACFIC are generally considered business elites, and they are normally leaders of privately owned, state-owned, or mixed ownership firms [59]. By contrast, the sector-based industrial associations have been developed outside the political system (tizhiwai), and these non-governmental associations usually follow the principles of "running independently, self-management, and self-financing." Associational membership embedded firms are more likely to adopt proactive ECSR practices. First, the industrial associations membership can prompt the proactive ECSR behavior of membership firms through the standards and norms of associations. That is, associational membership exerts normative institutional pressures on firms [14,60]. Because CSR has been the focus and one of the priorities of the work of ACFIC in the last decade, associational membership embedded firms are strongly motivated to practice proactive ECSR activities. For instance, to encourage and guide firms to fulfill CSR, the ACFIC released its first research report on the CSR of private firms in 2014 [61] and then released its first blue book on the performance of the Chinese private sector in fulfilling social responsibility in 2019 [62]. Apart from the ACFIC, sector-based industrial associations also attach great importance to CSR due to the rising pressures from consumers and the media [63,64]. The environmental irresponsibility of any membership firm will have a serious negative impact on the industry reputation. Due to the high potential costs caused by environmental irresponsibility (e.g., the loss of membership that will affect the exercise of other rights in the association) [65], associational membership embedded enterprises have the motivation to engage in proactive ECSR. In this regard, the engagement of embedded firms in proactive ECSR can be considered a response to the normative pressures.
In addition, industrial associations set guidelines on how the firms within the association should behave in environmentally responsible ways by promoting best practices [66]. Associational membership embedded firms may make effective proactive ECSR efforts due to excelling at accessing information about the industry best practices. Moreover, there are fierce homogeneous competitions among the numerous member firms. To distinguish themselves from peers and, thus, achieve competitive advantages, firms with membership would engage in proactive ECSR [15]. Forerunners will also provide imitation and demonstration effects for latecomers. That is, emulation could be triggered because potential ECSR adopters perceive diffused ECSR practices as appropriate [15]. In this sense, the pressures of associational membership embedded firms are mainly from peer firms, and the engagement of embedded firms in proactive ECSR can be regarded as a response to the cognitive pressures. Therefore, we propose the following hypothesis: Hypothesis 3. Associational membership embedded firms will make more proactive ECSR investments than non-embedded firms.

Family Involvement and Proactive ECSR
Many companies around the world are owned or controlled by families, and family business has been the world's most prevalent ownership pattern of business organizations [67]. For instance, La Porta et al. [68] documented that 53% of publicly-traded firms in 27 countries are family-controlled when using the 10% minimum control cutoff to judge a family business. Family-controlled firms also dominate the corporate landscape in China. For instance, 85% of private firms are owned by individual shareholders and their family (i.e., holding more than 50% of the shares) [69]. Scholars believe that family involvement (i.e., the power of families in shaping the goals, strategies and behaviors of firms; see [70]) has an influence on a range of firm behaviors (e.g., succession, governance, and CSR) [71], and quite a few studies have regarded family involvement as a key determinant of the ECSR activities of private firms [2,8,9,72,73]).
The socioemotional wealth (SEW) framework (see [74][75][76]) is a helpful theoretical tool that explains the behavior of family businesses. According to SEW theory, family firms are motivated by, and committed to, the preservation of their SEW, which is a primary reference point that family firms use in making major decisions [75,76]. The SEW refers to nonfinancial utilities that a family derives from its controlling position in a particular firm, such as the identification of the family with the firm, the ability to exercise family influence, and the perpetuation of family social capital and family dynasty [72,[74][75][76]. The basic assumption underlying SEW theory is that the willingness to preserve and enhance SEW is a unique utility with family involvement, and family principals are highly loss-averse to SEW [71]. Firms with higher family involvement are more likely to make decisions that are conducive to the preservation and enhancement of the SEW of their families, even when they contrast with firms' short-term economic goals [71,77].
Two dimensions of family involvement (i.e., family involvement in ownership and family involvement in management) can be distinguished on the basis of the existing literature [78][79][80][81]. Family involvement in ownership depicts the degree of family ownership and intrafamily ownership dispersion, while family involvement in management indicates the presence of the family members in the management teams of the company [78,79].
We contend that family involvement is conducive to proactive ECSR at each of the two dimensions. Specifically, when family owners have large shares in the business, they may become increasingly risk-averse and are sensitive about the external image they project to external stakeholders because they generally perceive the firms as an extension of their families [82]. Therefore, firms with higher family involvement in ownership are more likely to engage in proactive ECSR, which is conducive to the protection and enhancement of the family image and reputation, an important SEW goal [70]. Proactive ECSR is a discretionary responsibility that goes beyond the legal requirements [3,17], which means that firms can freely make decisions on how much ECSR investment to make despite the expectation of society that firms engage in proactive ECSR. Unlike reactive/passive ECSR (e.g., paying environmental pollution control fees in the context of strict government regulations), which generally aims to avoid imminent punishment, proactive ECSR aims primarily at accumulating reputational capital for firms and avoiding the potential damaging effect on the SEW of firms caused by being labeled as environmentally irresponsible corporate citizens [72]. Additionally, families holding large portions of shares have more legitimate power to influence the strategic decisions of firms, thereby gearing the strategic decisionmaking of firms toward family preferences [71], such as engaging in proactive ECSR. Therefore, families with higher family involvement in ownership are more motivated to engage in ECSR to avoid the potential risk of SEW loss.
Although family members tend to engage in proactive ECSR to accumulate and enhance the SEW of their family, they may be resisted by non-family members in their company. This is because non-family members pay more attention to financial values than non-financial affect-related values and fear proactive ECSR at the cost of the short-term economic profits of firms [10]. This is especially true in emerging markets, such as China [10]. Family involvement in management guarantees that the legitimate power granted by ownership can be employed directly to make decisions favoring family interests [71]. That is, firms can more easily implement strategic actions (e.g., proactive ECSR) that are conducive to preserving and enhancing the SEW of families. Furthermore, family involvement in management can strengthen the willingness family members to preserve the SEW of their family [71]. The more executive positions family members hold, the less clear the boundary between the family and the firm. In this sense, stakeholders are more inclined to perceive the firm as an extension of the family itself because the identity of the incumbent family executives is closely tied to the firm [83]. Thus, to protect the SEW of families, firms with higher family involvement in management tend to engage in proactive ECSR. We, therefore, hypothesize the following: Hypothesis 4. Firms with higher family involvement in ownership and management will make more proactive ECSR investments than firms with lower family involvement in ownership and management.

Firm Resource Base and Proactive ECSR
Firm resource base is considered a crucial factor influencing the engagement of firms in ECSR behaviors and ECSR investments (e.g., [17,[84][85][86]). Nevertheless, firm resource factors are normally included in the analysis in prior studies only as control variables. Given the importance of firm resource factors in predicting the ECSR efforts of firms, we regard them as focal determinants.
There are various ways to measure firm resource base, but firm size represents the most critical characteristic of firm resource base [87]. It is also a straightforward, observable, and, therefore, a solid practical measure of firm resource base [88]. Previous studies yielded inconsistent results as for the impact of firm size on firms' ECSR activities [29]. Given that the clear-cut distinction in size between large and small firms may be one reason of the inconsistent results, we only focus on large private firms and reason that, even for large firms, firm size is an essential factor influencing firms' engagement in proactive ECSR. Firms larger in size may be under more scrutiny of the public and the media due to higher visibility, and, thus, are more likely to engage in proactive ECSR to respond to the expectations of the public [4]. Moreover, firms larger in size generally have more capacity to undertake proactive ECSR because they are more likely to have slack resources (i.e., resources that are left unused in ordinary operational activities) [89] and, hence, are more responsive to the pressures and expectations of stakeholders.
The financial performance of firms is another factor that is commonly used as a proxy for firm resource base and a determinant for predicting ECSR (e.g., [4,84,85]). The available fund hypothesis believes that the CSR performance of firms is determined by its available financial resources, and higher levels of financial performance lead to higher levels of CSR performance [90]. That is, available financial resources increase a firm's ability to fund discretionary projects, such as proactive ECSR. Thus, firms with high financial performance are more likely to engage in proactive ECSR activities. The above arguments generate the following hypothesis: Hypothesis 5. Firms with a strong resource base will make more proactive ECSR investments than firms with a weak resource base.

Data
The data in this paper were collected from the 2014 Chinese Private Enterprise Survey (CPES), which was jointly conducted by four agencies, including the ACFIC, the State Administration for Industry and Commerce (SAIC), the China Society of Private Economy (CSPE), and the UFWD of the Central Committee of the CPC. The CPES survey has been conducted almost every two years since 1993, and hitherto, it has been conducted for 13 times [17]. As the data for 2016 and 2018 are unavailable, the data for 2014 are the newest data that have been open to researchers for academic use. The survey collects information from the previous year; thus, the firm data in this paper correspond to the year of 2013.
To be representative across regions and industries, the multiple-stage stratified random sampling method was employed to select sample firms among private firms in mainland China. A total of 6144 private firms were surveyed, which represent 0.04% of the total number of privately owned enterprises nationwide [91]. More importantly, the questionnaire includes information about firms' institutional embeddedness, family involvement, and proactive ECSR, which makes it one of the best datasets to explore the determinants of firms' ECSR practices in China. The CPES data have been widely utilized in prior studies published in top-tier academic journals, such as the Strategic Management Journal, Journal of Business Venturing, Journal of Business Ethics, and Journal of Comparative Economics (see [17,48,92,93]).
We excluded non-manufacturing firms considering that environmental pollution is mainly related to manufacturing firms. Given the peculiarity of environmental protection firms, they were also excluded. To verify, even for large firms, firm size is a critical factor affecting the proactive ECSR practices of firms; we only focused on large firms (firms with 1000 or more employees or firms with total sales of 400 million RMB and above) and, thus, eliminated small-and medium-sized firms (SMEs) (as for the criteria for judging SMEs, see China's Regulations on the Standards for Classification of Small-and Medium-sized Enterprises issued by four agencies, including the National Bureau of Statistics). The final sample includes 162 firm observations.

Method
Fuzzy-set qualitative comparative analysis (fsQCA) is used as a research approach in the present study. QCA was initially proposed by Charles Ragin (see [94]) to analyze datasets as a mid-way between a purely deductive variable-oriented methodology and a purely inductive case-based methodology [95]. Scholars believe that QCA has several advantages over traditional analysis methods (e.g., multiple regression analysis and structural equation modeling) [94,[96][97][98] and these advantages make the QCA method particularly relevant and mostly appropriate for this study.
First, different from traditional analysis methods that examine the net effects (i.e., direct and indirect effects) of individual independent variables on dependent variables, QCA aims at delineating the configurational effects of antecedent conditions on the outcome of interest [99]. The difference between net effects analysis and configurational analysis explains why QCA usually refers to causal conditions rather than independent variables. In configuration analysis, it is the joint presence or absence of a set of given causal conditions that lead to a particular outcome, while the notion of an independent variable means that it is possible to isolate the effect of such a variable on the dependent variable by changing its values when keeping all other independent variables constant [96]. A single causal condition is rarely sufficient to explain the presence of a particular outcome; on the contrary, outcomes are often produced by the conjunctural causation of multiple causal conditions [96]. Hence, QCA plays a greater role in dealing with complex causality by extending beyond the conventional interaction term effects in regression analysis, which are used to explain the interdependencies between independent variables, but tend to be constrained to three-way interaction effects [100]. Moreover, QCA enables the reduction of conditions for each pattern while seeking patterns of conditions that produce a specific outcome, so configurations only include necessary and sufficient conditions [101]. Existing studies have explored the impact of institutional embeddedness or family involvement factors on ECSR; however, some recent CSR studies (e.g., [16,83]) have shown that it is necessary to integrate institutional embeddedness factors with family involvement or firm resource base factors to understand the origin of the firms' CSR activities because these factors may coalesce to produce high CSR commitment. Since QCA can explore how key antecedent conditions coalesce to explain an outcome of interest, it is particularly suitable for this research.
Second, QCA explicitly allows for the equifinality of different combinations of causal conditions and stresses that the diverse configurations of causal conditions can lead to the same outcome [96]. The assumption of equifinality relates to the fact that "many roads lead to Rome." This advantage of QCA enables us to explore whether there are multiple causal recipes (combinations of causal conditions) leading to high proactive ECSR.
Third, QCA allows for the asymmetry of relationships to be captured and, thus, can identify different configurations of conditions that predict both the presence and the absence of a given outcome [99]. That is, the causal conditions leading to the presence of a given outcome do not need to be the inverse of the conditions that result in the absence of the same outcome. This type of asymmetry is difficult to capture by standard regression analysis [96]. QCA entails asymmetry, which makes it possible for us to explore different causal recipes that lead to the presence and absence of high proactive ECSR. This is important because the presence and absence of a given outcome may require different explanations.
Moreover, it is possible to keep the number of causal conditions small in QCA because it has no omitted variable bias in the same way a regression analysis does [96]. Specifically, in regression analysis, leaving out relevant independent variables not only decreases explanatory power, but also leads to biased coefficient estimates [102]. However, leaving out a relevant condition will decrease the explanatory power of the model, but will not produce an omitted variable bias because QCA draws on Boolean algebra rather than on correlations [96]. Since it is difficult to capture all possible factors influencing proactive ECSR, the QCA method is better due to its advantage in dealing with omitted variable problem.
Finally, QCA can work with relatively small sample sizes [103]. By contrast, traditional analysis methods, especially multiple regression analysis, require a relatively large sample size. Because the sample size of this study is not large, the advantage of QCA in dealing with small sample data is also one of the reasons for its application in this study.
Apart from fsQCA, the QCA technique includes several other variants, such as crispset QCA (csQCA) and multi-value QCA (mvQCA). Distinct QCA tools have different application conditions. In contrast with csQCA, which operates exclusively with dichotomous conditions, and mvQCA, which works with multi-valued conditions, fsQCA allows for establishing differences in degree [104]. Thus, we chose fsQCA as the methodological tool for this study. In fact, the fsQCA method has been increasingly used in CSR-related studies (e.g., [105][106][107]).
The fsQCA can be conducted in several steps. The first step is to calibrate the measures for the outcome and the causal factors into set membership scores, and fsQCA allows fuzzy membership scores, which range from 0 to 1 [108]. The second step involves building a truth table, which is a matrix with 2 k rows, where k is the number of causal conditions in the analysis [109]. The third step is to reduce the number of rows in the truth table considering both a frequency threshold and a consistency threshold [110]. In line with Navarro et al. [111], we specified a frequency threshold of 2 to rule out the less important configurations and a consistency threshold of 0.80, which is above the generally accepted minimum 0.75 threshold [103]. In the fourth step, we used the Quine-McCluskey algorithm, which is based on Boolean algebra to simplify the truth table. This algorithm provides three different solutions (a parsimonious solution, an intermediate solution, and a complex solution) through easy or difficult counterfactuals [110]. Two parameters were also employed to indicate the goodness of fit of the final solution, namely, coverage and consistency [109]. The fsQCA3.0 software was used to perform the fsQCA.

Outcome Measure
The outcome measure is the proactive ECSR of large private firms in China. Following Kim [3] and Zhou et al. [17], proactive ECSR refers to the voluntary environmental investment of firms for pollution abatement, such as sewage treatment, air pollution treatment, soil pollution treatment, solid waste treatment, flue gas treatment, and waste recycling. Specifically, proactive ECSR investment was measured as the monetary amount (10,000 yuan RMB) of a firms' voluntary investment to pollution abatement during the last year. In the sample, the mean proactive ECSR investment is 402.01 (10,000 yuan RMB), and approximately 85% (85.19%) of the firms engaged in proactive ECSR.

Causal Conditions
Following existing studies [6,[14][15][16]27], we employed three dichotomous variables to measure the institutional embeddedness of firms. Specifically, political membership embeddedness was measured by the top manager's membership of the NPC or CPPCC (1 = membership in any of the two parties; 0 = no membership); party branch embeddedness was gauged by assessing whether there is a party branch of the CPC within the firm (1 = yes; 0 = no); and associational membership embeddedness was measured by the top manager's membership of the ACFIC (1 = membership in the ACFIC; 0 = no membership). Given the data availability, the associational membership embeddedness did not involve sector-based industrial associations (hangyexiehui).
As mentioned before, family involvement was measured from two dimensions: family ownership involvement and family management involvement. Following prior literature [71,78,80], family ownership involvement was gauged by the percentage of the ownership held by the family members. Referring to prior research [70,82,112], family management involvement was measured on the basis of whether the major decisions of an enterprise are made by the main investor himself or the position of the chairman of the board is held by the main investor or his family members in the case of having a board of directors (BoD) (coded as 1 if any of the two questions were answered in the affirmative, otherwise as 0).
Following the existing studies [4,85,87,88], firm size and financial performance were utilized as proxies for firm resource base. In line with the existing studies [71,78], the firm size was measured as the logarithm of the total sales (10,000 yuan RMB) in 2013, which is a common measure for firm size. Following previous studies [4,16,113], the firm profitability was used to measure the financial performance of firms, and it was measured by the ratio of net profits over sales (ROS).

Calibration
Before building the truth table, fsQCA requires a calibration for variables, namely, the transformation of variables into sets according to their degree of membership to a specific condition. The scores after calibration range from 1 (full membership) to 0 (full non-membership). Following the suggestion of Ragin [114], this study employed a direct calibration method, which uses three anchors to calibrate continuous measure values into membership values. The anchors include the thresholds for full membership (0.95), full non-membership (0.05), and the crossover point of maximum ambiguity (0.5). Following the existing studies [99,108,115], we used the 10th, 50th, and 90th percentiles as the full nonmembership, crossover point, and full membership thresholds, respectively. Considering the relatively discrete distribution of proactive ECSR investment (see its standard deviation in Table 1), we used the 20th, 50th, and 80th percentiles as the full non-membership, crossover point, and full membership thresholds, respectively, following the suggestion of Fiss [95]. Regarding dummy variables, the calibration for them can be performed directly from the original variable into a crisp set, where 1 indicates full membership and 0 specifies full non-membership [116]. Table 1 shows the descriptive statistics and set calibration thresholds.

Analysis of Necessary Conditions
Given that QCA assumes complex causality, it requires the analysis of necessary conditions to determine whether any causal condition can be considered a necessary condition for the specific outcome, which, in this study, is proactive ECSR investment. According to Schneider and Wagemann [117], a condition can be considered necessary when its consistency score is greater than 0.9. Table 2 displays the analysis results of necessary conditions considering both the presence or the absence of the antecedent conditions. Table 2 exhibits that the presence of associational membership embeddedness turns out to be a necessary condition. Moreover, the presence of family management involvement can also be labeled "necessary" because its consistency score is over the threshold. That is, the presence of associational membership embeddedness and family management involvement makes a strongly stand-alone influence on the outcome of interest (high proactive ECSR investment), respectively. As the consistency value among any other conditions is below 0.9, none of any other conditions alone is a necessary condition for the outcome high proactive ECSR investment. This highlights the necessity of QCA to further explore the causal contribution of the different configurations of conditions to the given outcome.  Table 3 presents the results of the fsQCA regarding the influence of seven conditions on proactive ECSR investment. Following the notation introduced by Ragin [103], blackfilled circles indicate the presence of a condition, white crossed-out circles represent its absence, and blank spaces indicate that it is irrelevant whether the condition is present or absent. The fsQCA yields three possible solutions: complex, parsimonious, and intermediate solutions. Among these, the intermediate solution is generally considered the best one, and because the complex solution avoids using any counterfactual cases [118], it is too complex to interpret [119]. The parsimonious solution allows for the use of any remainder that will generate simpler or fewer recipes [118], thereby blurring the boundaries between theories and cases [119]. Therefore, we use the intermediate solution for interpretation. Following Fiss [95], the core and peripheral conditions are distinguished by employing parsimonious and intermediate solutions. A core condition is the condition that appears in both parsimonious and intermediate solutions, and it has a strong causal relationship with the outcome. By contrast, a peripheral condition is the condition appearing only in the intermediate solution, having a weak causal relationship with the outcome.

Analysis of Sufficient Conditions
For each configuration of causal conditions, Table 3 presents the raw and unique coverage and consistency. The raw coverage indicates the proportion that a configuration accounts for the outcome cases, while unique coverage denotes the proportion that a configuration uniquely covers the outcome cases [103]. The consistency indicates the extent of cases corresponding to the set theoretic relationships expressed in a solution [95].
In addition, Table 3 shows the solution consistency and solution coverage: the former measures the degree to which all solutions together consistently result in the outcome, and the latter implies the extent to which the extracted set of solutions accounts for the outcome of interest [101]. According to Eng and Woodside [120], a model is informative when its consistency value is greater than 0.74. Table 3 exhibits that the solution consistency is 0.833, which can be considered a good fit parameter. Moreover, the solution coverage is 0.154, which indicates that the causal configurations in Table 3 account for 15.4% of the empirical evidence. Table 3 also depicts that two causal paths can lead to high proactive ECSR investment. Although they are different, both of them are sufficient conditions for achieving high proactive ECSR investment, thereby highlighting the necessity of QCA in analyzing the origin of proactive ECSR because QCA allows for the equifinality of the divergent configurations of causal conditions that can lead to the same outcome.
As the fsQCA results show, no condition alone is sufficient to account for high proactive ECSR investment, which means that the configurations of institutional embeddedness, family involvement, and resource base factors of firms are associated with high proactive ECSR investment rather than these factors independently. The existing research only separately analyzes the influence of certain factors (institutional embeddedness or family involvement factors) on the proactive ECSR of firms, thereby failing to capture the antecedents of the proactive ECSR of private firms comprehensively and thoroughly. Furthermore, analyzing the antecedents of proactive ECSR by using the QCA method is quite necessary given that traditional quantitative analysis techniques could not effectively account for complex causality.
Configuration 1 suggests that a combination of political membership embeddedness, associational membership embeddedness, family management involvement, large firm size, and the absence of party branch embeddedness goes together with the high proactive ECSR investment of large firms. Moreover, for firms in Configuration 1, family ownership involvement and high financial performance are irrelevant.
Configuration 2 indicates that the joint influence of party branch embeddedness, associational membership embeddedness, family management involvement, and large firm size leads to the high proactive ECSR investment of large firms, although the absence of political membership embeddedness, family ownership involvement, and the negation of high financial performance accurately exists in this particular situation.
Interestingly, the presence of associational membership embeddedness and family management involvement is always associated with large firm size to produce high proactive ECSR investment and, therefore, appears to provide complementary effects on the proactive ECSR investment of large firms. In the meantime, these three conditions, combined with political membership embeddedness or party branch embeddedness, can drive the high proactive ECSR investment of large firms. Hence, political membership embeddedness and party branch embeddedness can be perceived as substitutes. Political membership embeddedness, thus, decreases the effectiveness of party branch embeddedness in achieving the high proactive ECSR investment of large firms.
As we can see from configurations 1 and 2, the presence of family management involvement has an important influence on the proactive ECSR investment of large firms. By contrast, a high level of family ownership involvement seems irrelevant, which may be because this study targets large private firms and they generally have a relatively perfect corporate governance mechanism (e.g., the presence of the BoD as the strategic decisionmaking group) [121]. Under this circumstance, family management involvement plays a greater and more direct role in inducing large firms' proactive ECSR activities than family ownership involvement. This highlights the necessity of subdividing family involvement to several components to more accurately identify the factors that can drive the proactive ECSR of firms.
The results indicate that the resource base factors of firms are substantially important, resulting in high proactive ECSR investment even for large firms. However, firm size exerts a considerable influence on large firms' proactive ECSR activities, while the effects of financial performance on proactive ECSR can be ignored. This may be due to the fact that the engagement of proactive ECSR needs more than financial support (especially support from short-term financial performance) and firm size can capture the resource strength of firms more comprehensively.

Robustness Checks
Several robustness checks were performed as recommended by Skaaning [122] to verify the stability of the fsQCA results. First, changing the consistency thresholds is one of the common practices for robustness checks [122,123]. Following Farivar and Richardson [124], the fsQCA was repeated with a reduced consistency threshold of 0.75, which is considered the minimum recommended threshold [103]. The findings validated that there were still two configurations achieving sufficient consistency and that the solution consistency did not decrease radically (Table 4). Moreover, the solution coverage increased negligibly to 0.167, and the elements of configurations 1 and 2 in Table 4 are almost the same as those in Table 3. Hence, the findings are robust to the use of the alternative specification of a consistency threshold. Second, following An et al. [123], we changed the calibration thresholds for the outcome from the 20th, 50th, and 80th percentiles to the 10th, 50th, and 90th percentiles and repeated the analysis of the sufficient conditions of high proactive ECSR investment. As is shown in Table 5, we found that the new results are consistent with the original results; hence, the stability of the findings was further confirmed. Finally, following the suggestion of An et al. [123], we conducted a sufficiency analysis with the negation of proactive ECSR investment as an outcome (Table 6). It is believed that if the combinations of conditions explaining an outcome also explain the negation of the outcome, then such ambiguous combinations should not be considered sufficient solutions [108]. As is shown in Table 6, the sufficient configuration leading to high proactive ECSR investment and the sufficient configuration resulting in non-high proactive ECSR investment are asymmetric. That is, no configuration is relevant for both the presence and the absence of the given outcome. Hence, the aforementioned findings remain unchanged. In sum, we can conclude that the fsQCA results were robust. It is also worth noting that even when there is high involvement of family ownership, a small firm size and the absence of party branch embeddedness can drive the non-high proactive ECSR investment of large firms.

Main Findings and Implications
Engaging in ECSR is becoming an increasingly common business practice globally; moreover, ECSR is increasingly becoming a topic of interest among researchers. Although many studies have investigated the determinants of ECSR from the perspective of either institutional embeddedness or family involvement, to the best of our knowledge, this is the first study that explores the configuration of causal conditions that lead to the high proactive ECSR investment of large firms by simultaneously considering institutional embeddedness, family involvement, and the resource base factors of firms.
In sum, the findings verify that although associational membership embeddedness and family management involvement can be labeled as necessary conditions and make a strongly stand-alone influence on the high proactive ECSR investment of large firms, respectively, no condition alone is sufficient to account for high proactive ECSR investment. Institutional embeddedness, family involvement, and the resource base factors of firms must interact with one another to create one or more best-fit configurations that are associated with the high proactive ECSR investment of large firms. Hence, this study highlights the importance of considering factors concerning institutional embeddedness, family involvement, and firm resource base in understanding the driving forces of proactive ECSR and, thus, echoes the recent studies that have integrated institutional embeddedness factors with family involvement or resource base factors of firms to understand the origin of the CSR activities of firms [16,83].
The findings further confirm that there is more than one configuration of causal conditions that can achieve high proactive ECSR investment. Moreover, the findings validate that it is not the presence of all institutional embeddedness, family involvement, and firm resource base factors that results in high proactive ECSR; rather, the specific combinations of components of three types of factors can be sufficient for high proactive ECSR. Interestingly, the study argues that associational membership embeddedness, family management involvement, and large firm size are always combined for high proactive ECSR, confirming their complementary relation. As such, the findings of this study deepen the understanding of the determinants of the ECSR activities of firms.
As noted, we observe that the presence of these three factors (associational membership embeddedness, family management involvement, and large firm size) in combination with political membership embeddedness or party branch embeddedness can drive high proactive ECSR investment, indicating that political membership embeddedness and party branch embeddedness can be perceived as substitutes. Apart from "LT" that stresses maintaining legitimacy is the driver of the engagement of firms in CSR activities, some scholars have adopted the social exchange theory (reciprocity logic) to explain the effect of party branch embeddedness on the engagement of firms in CSR (see [17]). According to this theory, party branch embeddedness can bring benefits to enterprises, and firms' participation in CSR can be viewed as feedback. However, the benefits brought by party branch embeddedness are not intuitive. Given that political membership embeddedness and party branch embeddedness are substitutes for each other, their effect and mechanism on the proactive ECSR of firms should be similar, thereby verifying at least to some extent the applicability of "social exchange theory," because researchers generally believe that political membership embeddedness can bring firms certain benefits [6,27].
The findings assert that family management involvement has an important influence on the proactive ECSR investment large firms, while family ownership involvement seems irrelevant. On the one hand, this study confirms the research finding of prior studies that family involvement is an essential determinant of firms' CSR activities [2,82,125] and that the ability to exercise family influence is a crucial dimension for understanding the proactive ECSR of firms with a high level of family involvement among all the SEW dimensions. On the other hand, it elucidates the importance of subdividing family involvement components to more accurately identify the component that can drive the proactive ECSR of firms and echoes the call to provide a more nuanced understanding of the family involvement-CSR nexus [71,78].
This study demonstrates that the resource base factors of firms are substantially important in driving high proactive ECSR even for large firms. Hence, it can be inferred that the role of resource base factors of firms may be greater in determining the extent to which large firms and SMEs engage in proactive ECSR. In this sense, it may be not appropriate to regard the resource base factors of firms just as control variables in previous studies, and renewing our recognition of the importance of firm resource base factors in determining firms' ECSR is fairly indispensable.
The findings of this study have critical managerial implications. First, these findings confirm that the ongoing construction of the CPC organizations at the grassroots level has objectively promoted private firms' engagement of proactive ECSR to some extent. In this sense, to promote the engagement of private firms in proactive ECSR, the authorities concerned could encourage firms (especially non-political membership embedding firms) to establish a CPC party cell within their firms. Second, the findings inform policy makers that it is difficult to obtain the desired result by solely relying on public policy propaganda to promote the proactive ECSR investment of firms because small-scale firms may lack sufficient resources to conduct ECSR activities. Hence, policy makers should encourage large-scale firms to actively fulfill ECSR as a priority. In sum, the research findings could provide references for the design and implementation of policies aiming at promoting the ECSR activities of firms. For family business owners, the findings imply that family management involvement plays a greater role in determining proactive ECSR than family ownership involvement; thus, the controlling family of a family firm should take an advantageous management position within the firm to implement proactive ECSR practices to build firm advantage.

Limitations and Directions for Future Research
This study is not exempt from limitations. First, we only focused on large firms in China; therefore, the applicability of our conclusions to SMEs or firms in other countries needs further investigation. Second, we examined the main institutional embeddedness, family involvement, and resource base factors rather than investigating all possible factors. Further research could investigate the effects of other factors concerning institutional embeddedness, family involvement, and firm resource base, such as the duration of family control, which is considered one of the dimensions of family involvement by few scholars (see [71]). Finally, the limitations of the QCA method must be acknowledged. Although the QCA method has distinct advantages over traditional analysis methods, it also has several limitations. First, QCA typically tests for sufficiency, not for necessity, and although fsQCA has the capacity to examine necessity, this is done separately from the sufficiency analysis [96]. Second, the results produced by QCA become increasingly complex as the number of antecedent conditions increases and, hence, are difficult to interpret [96]. Last but not least, the fsQCA method is sensitive to set calibration [109]. Nevertheless, we have tried to confirm the stability of research conclusions by performing several robustness tests. Research that can provide clear guidelines on how to calibrate fuzzy sets in CSRrelated research was warmly encouraged, which could significantly improve research comparability. In addition, to comprehensively and thoroughly explore the antecedents of proactive ECSR, studies that integrate structural equation modeling and QCA methods and utilize specialized enterprise survey data (rather than general enterprise survey data) were warmly welcomed.

Conclusions
What is the origin of the proactive ECSR of firms? Is the proactive ECSR of firms driven by institutional embeddedness factors or promoted by family involvement factors or dependent on the resource base factors of firms? Although there has been growing interest in understanding how institutional embeddedness or family involvement factors contribute to ECSR in the existing CSR literature, the cross-fertilization of this literature has been limited. QCA allowed us to analyze complex interaction effects between these factors on the proactive ECSR of firms. In this study, we sought to advance the ECSR determinant literature by suggesting a configuration way of thinking about these antecedents using the fsQCA method. The findings provide new insights into how institutional embeddedness, family involvement, and resource base factors of firms are combined to produce high proactive ECSR. Particularly, the results verify that individual factors cannot lead to high proactive ECSR by itself, and the appropriate configuration of these factors drives the proactive ECSR of large firms. Knowledge of the factors that affect firms' ECSR is one of the key research questions to be solved among academics, and this study provides a meaningful starting point for future research by highlighting the necessity of integrating institutional embeddedness, family involvement, and the resource base factors of firms.