Is Corporate Social Responsibility Used to Mask Corporate Speculation? Evidence from Emerging China

Abstract

Why do rational, profit-orientated firms generously engage in corporate social responsibility (CSR)? Our study explores the real motives of speculative firms for CSR engagement and the hidden causality behind it. Using national survey data of privately owned firms in China, we find that corporate speculation positively influences firms’ engagement in CSR, revealing that CSR is instrumental and that firms use it as a tool to mask their speculative activities by building their reputations and buying ‘leniency insurance’ against potential penalties. Further, the relationship between speculation and CSR is less pronounced in firms with political involvement, revealing that the effect of political involvement as an informal institution somewhat protects speculators from potential sanctions without a CSR premium. We also discovered that the relationship between corporate speculation and CSR—as well as the moderating role of political involvement—is less pronounced among developed regions, revealing that the development of formal institutions can restrict the instrumentality of CSR and the effect of political involvement. Such findings have important implications for CSR in emerging economies.

1. Introduction

Our study explores whether and how corporate social responsibility (CSR) can be used instrumentally to mask corporate speculation in emerging China, and further identifies the mechanism as well as the boundary conditions.

CSR engagement is widely employed as a prevalent and necessary approach to maintain and improve a firm’s relationship with society. An increasing number of studies of CSR in emerging economies have explored corporate philanthropy, among many other major components of Carroll’s pyramid of social responsibility [1,2], as a primary and direct measure of CSR [3,4,5,6,7,8]. This is especially true in China [9,10,11]. With the Wenchuan earthquake in 2008 as a milestone, corporate philanthropy has long become a dominant indicator for corporate social responsibility in this country [12]. Since then, annual philanthropic donations from firms have contributed more than half of China’s philanthropy. In 2017, national statistics showed that Chinese firms donated about 96.3 billion RMB (approximately $13.9 billion USD) to charitable causes, accounting for more than 64% of the total national philanthropy. Donations from privately owned firms reached 48.2 billion RMB (approximately $7.1 billion USD), accounting for 51.1% of the total corporate philanthropy, and more than 60% of the firms that donated more than 0.1 billion RMB were private firms [13].

Firms’ reasons for engaging in socially responsible activities range from purely altruistic (benefiting others) to purely instrumental (benefiting the firm itself) [14,15,16,17]. However, some scholars suggest that studies focusing on emerging economies should pay more attention to the extensive instrumentality of CSR, which has become increasingly prevalent in recent years, because it is hard to find purely altruistic CSR in relevant countries [18,19], and even the altruistic explanation of CSR in some cases may be modified by one or more instrumental explanations [15,20]. From an instrumental perspective, CSR is seen as a pragmatic tool to obtain both monetary and non-monetary returns [19,21], including improving financial performance [22], responding to stakeholder expectations for societal acceptance [21,23,24] and satisfying the public demands to avoid potential boycotts [25,26].

In addition to the above-mentioned instrumental motives, an increasing number of studies have argued that CSR can be used to manage or reduce risks by generating moral capital and providing insurance-like protection [18,27,28]. Several studies have indicated that CSR is an important investment for reputation-building and maintenance [23,24]. Some recent studies have further demonstrated that firms may simply use CSR to whitewash corporate irresponsibility [18,29], in which unfortunate misconduct by the firm or its officers might be offset by previously accumulated goodwill and reputation [30,31,32]. Although these studies have enriched our understanding of the instrumentality of CSR engagement by exploring its risk-managing roles, much attention has been paid to reputation-building or misconduct offsetting, while other hidden causalities—such as profit-sharing/exchange—has been underexamined. For example, whether and how CSR engagement brings insurance-like benefits to privately owned firms by providing protection against possible sanctions in emerging economies where the relationship-based assets are greatly valued has been largely ignored in extant literature.

Our study attempts to fill this gap in the literature by placing the research in the emerging context of China and focusing on two common—but contradictory—behaviors of privately owned firms in China:—corporate speculation and CSR engagement—and exploring whether CSR is used by speculators to exchange moral capital and reduce negative public pressure and thus buy insurance-like protection from salient stakeholders. We focus on Chinese privately owned firms for two primary reasons. First, environmental uncertainty and complexity in emerging China result in rational and speculation-orientated privately owned firms [33,34,35,36] that seek short-term efficiency and make easy money by engaging in activities unrelated to their main business. In searching for the reason that rational, profit-orientated firms generously engage in CSR, we were inspired to inquire as to their real motives for CSR engagement, and such seemingly contradictory aspects of corporate behavior are suitable for us to explore the potential causality and hidden relationships between corporate speculation and CSR. Second, China’s rapid economic reform has shaped or even created a unique social, economic and cultural environment that is quite different from those in rule-based economies [30,37,38]. All of these special environmental factors and their interactions [39] have provided a unique setting for researching the hidden causality of CSR engagement. Moreover, regional economic development disparities have allowed us to determine the changes in the focal relationship and causality and thus better understand the mechanism.

Using a nationwide survey of privately owned firms, we report several important findings. First, we found that corporate speculation was significantly positively correlated with CSR engagement. Following the instrumental CSR view, speculative firms use CSR as a risk-managing approach to mask speculative activities and buy ‘leniency insurance’ [18,27,28] from key stakeholders. Second, the positive connection between corporate speculation and CSR engagement was less pronounced in firms with political involvement. This finding suggests that insurance-like protection arising from CSR will be substituted quickly by establishing direct political protection. Third, the above mechanisms vary across regions with different levels of social and economic development, suggesting that the instrumentality of CSR will diminish gradually with economic development. All the main results of our study hold in various robustness checks.

Our study contributes to the literature in three ways. First, it is one of the few studies to explore the hidden connection between corporate speculation and CSR engagement. We provide systematic evidence to confirm that managing or reducing risk is an underestimated instrumental motive for CSR engagement. Our study enriches the instrumentality of CSR in the context of emerging countries and extends the growing discussion on the relationship between business and society [40,41].

Second, our study illustrates the moderating role of informal institutions on the link between corporate speculation and CSR. This finding enriches our understanding of the focal relationship by demonstrating the substitutive effect of political involvement on moral capital as well as insurance-like protection. Thus, our study echoes the existing literature that informal institutions exert a strong influence on firms’ social engagement [37,42], and extends CSR research by uncovering the hidden causality in emerging economy settings and by further exploring the instrumentality of CSR.

Third, we find that the level of regional development moderates the relationship between corporate speculation, political involvement and CSR engagement. This finding demonstrates that the instrumentality of CSR is shaped by regional, social and economic heterogeneity. Thus, our study fills a gap in the current literature by identifying rarely tested boundary conditions within the context of China.

The structure of the paper is as follows: in the Section 2, we discuss and review the background and related literature, then develop our research hypotheses. In Section 3, we illustrate the research methodology, including the sample, empirical models and variables. In Section 4, we report the empirical results of descriptive statistics, correlation analysis and multivariate analyses. Finally, we summarize our conclusions, discuss the contributions and present the implications.

2. Research Context, Literature Review and Hypotheses Development

2.1. CSR and Instrumental Motives

Although definitions vary considerably within the extant literature, CSR in our study refers to firms’ social concerns in their business operations and in interactions with their stakeholders [43]. Thus, with the absence of uniform standards, firms are more likely in practice to be involved in social activities with high acceptability and identification among stakeholders [44], resulting in philanthropic contributions as a safe option [4,5,6,7,8]. Therefore, in emerging economies, it is relatively easy to find firms that participate in corporate philanthropy—the most discretionary and crucial dimension of Carroll’s pyramid of social responsibility—as the proxy and direct measure of CSR, as exemplified in China [9,10,11].

A significant body of literature has identified the motives for firms’ engagements in CSR and in corporate philanthropy, ranging from purely altruistic to purely instrumental [14,15,16]. However, from an instrumental perspective, scholars insist that CSR is a pragmatic tool used to complete the task of developing a ‘business case for CSR’ [17,45]. That means that a firm engages in CSR activities only if doing so provides it with valuable benefits. The instrumental motives for CSR in extant literature have been classified as (1) improving financial performance, (2) responding to stakeholder expectations for operations and (3) complying with public demands to avoid potential boycotts.

Hence, CSR is instrumentally motivated if firms focus on monetary returns from CSR activities. The connection between CSR and corporate financial performance has attracted much more attention than any other topic. Although there is no consensus, an increasing number of scholars and practitioners have accepted that engagement in CSR has a positive but mild effect on financial performance [46]. CSR can also be viewed as a strategic tool for achieving market competitiveness [47] or for improving firms’ attractiveness to potential employees [43]. CSR is instrumentally motivated when firms invest their resources to establish a connection with salient stakeholders [21,23,24], mainly to obtain societal acceptance and generate a good image. CSR engagement may help firms establish relationship-based assets with the local community, officials and regulators [48] for their business operations. CSR is also considered to be instrumentally motivated if firms’ social involvement is in accordance with demands from the public—especially in regard to non-government organizations [25,26]. CSR engagement may generate non-monetary benefits, such as efficiency gains from improved access to resources and cost savings from improved contact with regulators.

In addition, a growing branch of study has attempted to demonstrate the instrumentality of CSR engagement from a risk-management perspective [18]. Previous studies have supported this view by indicating that CSR is used by firms to build or repair their reputations [23,24] and thus alleviate pressure from stakeholders. Some studies have found that firms with poorer social performance were more likely to make philanthropic donations than their counterparts with better social performance [25]. Du (2015) showed that CSR engagement has been used as a kind of moral window-dressing to overshadow firms’ wrongdoings and divert public suspicions [30]. Moreover, many researchers have argued that firms use CSR to offset bad perceptions as well as to buy good will or a better reputation [29]. Although such studies, by focusing on irresponsible or illegal firm behaviors, have advanced our knowledge about the instrumental roles of CSR in mitigating a firm’s misconduct, little attention has been paid to the hidden causality between normal or seemingly normal firm behaviors and CSR engagement.

Furthermore, some theoretical studies have shown that CSR engagement has been able to generate moral capital and thus provide insurance-like protection for corporate operations in an uncertain environment [18,27,49]. Such insurance-like protection refers to the ways in which moral capital from CSR engagement can service as a buffer against negative influence and reputation risk that may arise in the course of business [50]. Hence, all the expenses of engaging in CSR can be regarded as an insurance premium that a firm pays in exchange for leniency from key stakeholders and protection against negative consequences [18]. Some studies have supported such a view by examining the insurance-like effect of CSR on a firm’s financial performance in terms of market value [27], shareholder value [18] and stock and bond prices [49]. However, whether and how CSR engagement provides insurance-like benefits for firms in emerging economies has been underexamined.

Therefore, our study attempts to fill the gaps discussed above by studying CSR in the emerging context of China and focusing on two common—but contradictory—behaviors of privately owned firms in China (i.e., corporate speculation and CSR engagement), examining whether CSR is used by speculators to generate moral capital and then obtain insurance-like protection from key stakeholders.

2.2. Corporate Speculation in Emerging China

Since 1978, China has been transforming its planned economy, under strict central government control, into a socialist market economy [51]. Following that, privately owned firms have emerged and developed rapidly. In recent years, these firms have contributed more than 60% of GDP and more than 80% of job opportunities in China, showing that they have aligned with state-owned enterprises as a dominant business form in the biggest emerging economy [52,53].

However, many private firms in China are struggling for survival and growth. For a long time in China, some industrial practices and policies have explicitly or implicitly ‘discriminated’ against private firms by favoring state-owned enterprises [10,54]. For instance, there are still ‘glass ceilings’ for private firms in industries that are traditionally controlled or even occupied by state-owned monopolies. Governments at all levels often take control of various resources and are usually more supportive of state-owned enterprises [52,54], while private firms are sometimes neglected. Unfair competition and unfavorable institutional conditions have resulted in more-severe resource limitations and growth liabilities for private firms, and have driven them to pursue short-term efficiency. Therefore, corporate speculation—such as investing in real estate, the stock market, the futures market or private lending—has become an easy choice for many private firms, which usually brings short-term economic benefits and thus seems quite rational.

Moreover, economic transition is always accompanied with increasing uncertainties, and in transitional China the social and economic environment has been characterized by such uncertainties, which makes it difficult for firms—especially those which are privately owned—to predict and control their own futures. Thus, corporate speculation has become a common and accepted part of doing business in China [55]. After the economic crisis in 2008, the profitability of many firms declined sharply due to the severe economic recession, and firms may have found themselves even more reliable on speculative investments to make money quickly and to maintain balance. As a result, many private firms engage in activities unrelated to their main business, and speculative behaviors that target short-term profits have become increasingly popular in China [34,35,56,57].

Overall, the speculation of privately owned firms in China refers to “self-interest seeking with guile” [58] (p. 47), with goals to improve a firm’s performance by making money quickly [59,60]. In China, there are some speculative activities that are favored by many private firms [35,36,57], and we can roughly divide them into three categories. The first type is investing in areas that are not related to a firm’s main business. For instance, China’s increasingly high housing prices have made investing in real estate a profitable choice for many private firms. The second type is directly investing in the stock market, which clearly shows corporate speculation as well as a firm’s desire for short-term economic benefits. The third type is investing in private lending with high returns. Such speculative investments improve firms’ short-term profitability and adaptability within an uncertain environment, and thus seem quite rational in the short term; however, in the long run, such speculative behaviors can result in negative consequences that may be gradually or suddenly revealed, and which can damage sustainability [35]. For instance, in China, speculative investments have directly contributed to the long-lasting bear market since 2008, and the boom in the real estate industry since 2015; both are now under much tighter regulatory control of central and local governments. Also, in recent years, many private lending businesses have been exposed to disruptions in the capital chain, and some of them have even suddenly disappeared, leaving investors with irredeemable losses.

In general, corporate speculation may damage a firm from many perspectives. First, speculative activities—which typically have high short-term returns—actually put a firm at high financial risk [34], and failure of such activities may directly and greatly damage firm performance. Second, speculative investments usually take some resources for innovation and business expansion, and thus they damage firm growth and shareholder interests in the long run [36]. Third, long and frequent involvement in speculation is very likely to be criticized by stakeholders, and does harm to a firm’s reputation and image. Finally, some speculative activities, such as investing in real estate, are likely to cause economic bubbles and social problems, while others, such as private lending, are not recognized or supported by governments [34,57]. Hence, corporate speculation can negatively influence a firm’s relationship with officials and regulators, and having positive relationships with these individuals is crucial for private firms in China.

Considering such possible negative consequences, it is not difficult to infer that corporate speculation clearly reveals a firm’s desire for short-term economic benefits. However, many private firms, while active in speculative investments, are willing to give up some economic benefits for CSR activities. These self-serving private firms have long played a prominent role in China’s social and charitable causes. Although some previous studies have argued that corporate philanthropy in Chinese private firms is a window-dressing activity [29,30], it is not easy to deny private firms’ generous contributions to society, which is more or less contradictory to their generous investment in speculative activities.

Thus, we are motivated by the current statuses of privately owned firms in emerging China to study the relationship between speculation and CSR engagement.

2.3. Hypotheses Development

2.3.1. Corporate Speculation and CSR Engagement

Since all firms operate within a wider social system, their activities are normalized by many salient stakeholders in the embedded environment, including investors, employees, suppliers, partners, customers, communities and regulators [48,61]. Thus, firms assume social responsibility and act philanthropically in line with social demands and stakeholder expectations for several reasons [21,23,24]. For rational and self-serving privately owned firms, CSR engagement can be used as a risk-managing tool to generate goodwill and mitigate stakeholders’ negative assessments toward business practices and even wrongdoings [18,29,30]. Moreover, engaging in continuous and long-term social activities may reduce potential negative impact on firm performance, such as market value [50], stock price and bond price [49], especially in the face of certain negative events.

Hence, CSR engagement helps private firms generate a heap of positive appraisal to mask their speculative orientation and divert public attention from their self-serving activities. In other words, firms use CSR to share their profits, especially those from speculative activities, with stakeholders and other people in society, and thus mitigate the negative image of speculation or myopia. Such seemingly generous ‘profit-sharing’ models may lead to favorable public responses, bringing firms social recognition and positive reputations [28,29]. One of the most glaring examples is Jia Duo Bao, a Chinese herbal soft drink manufacturer which quickly became one of China’s most well-known and highly esteemed brands after a timely and generous donation of 100 million RMB (about 14.5 million US dollars) right after Sichuan earthquake in May 2008. Despite its positive image in the real estate industry, Vanke suffered severe public criticism and backlash as a response to its chairman’s initial pledge of only 2 million RMB (about 0.28 million US dollars), and was considered a miserly ‘iron rooster’, which negatively influenced its brand image and stock price after the earthquake [12].

Furthermore, as soon as salient stakeholders receive the profit-sharing signal of CSR from a certain private firm, it is relatively easy for the firm to acquire and accumulate moral capital that can temper skepticism from stakeholders if the firm is given a negative assessment in the future [18,49]. Thus, CSR engagement acts as a type of insurance-like protection for firm operation [27,28,49]. Firms’ ongoing social engagement creates a positive image and convinces the public that most of their actions are done with good intentions. Thus, even if they are found engaging in unethical speculative behaviors, stakeholders and society tend to believe that such behaviors are free of any evil intentions and thus forgivable [18,49]. Speculative firms, who package themselves as good citizens through CSR, are able to divert the public’s attention from suspicious or improper earnings by pre-purchasing good reputation and moral capital. CSR forms a protective buffer between firm behaviors and the expectations of sympathetic stakeholders [6], and thus enables them to survive possible negative publicity or even scandals.

More importantly, CSR can be used as an effective tool to ‘buy off’ regulators to obtain ‘leniency insurance’ against potential penalties [37,49]. In some Western countries, some firms are able to openly seek political rents via political donations and lobbying, but there is no such institutionalized rent-seeking approach in China. Realizing that local governments are usually in need of resources for social causes, firms find that socially responsible behaviors, especially donations to address governmental concerns such as education, poverty-alleviation, medical development and disaster relief are effective in winning governmental recognition [30,37]. To some extent, contributions to such politically preferential causes can be regarded as a political investment [18] that secures favorable regulatory conditions [62] and mitigates some possible political uncertainties [63]. Speculative firms, driven by their need for such political support, tend to be more active in rent-seeking CSR activities, in the hope of concealing and mitigating the negative effects of unsustainable behaviors [10,35].

Therefore, based on the above, we can assume that:

Hypothesis 1.

Corporate speculation positively influences CSR engagement.

2.3.2. The Substitutive Role of Informal Institutions

A prominent feature of emerging economies [64,65] is that formal institutions and legal systems are not fully developed, allowing informal institutions to play a more prominent role in business. In particular, political involvement serves as a powerful informal institution. Firms usually build, maintain and adjust their political networks to deal with institutional uncertainty and then acquire resources that are not available in the market [51,66]. This study conjectures that firms’ political involvements influence the relationship between corporate speculation and CSR for the following two reasons.

First, political involvement has a direct connection with CSR engagement [37,63,67]. Since “social responsibilities of businessman arise from the amount of social power that they have” [68] (p. 48), a firm is very likely to lose its position in society if it does not use its social power responsibly in accordance with a society’s demands [69]. Therefore, CSR engagement, especially corporate philanthropy, becomes an ideal way for firms to reciprocate their social power. As has been pointed out by some scholars [70,71], Chinese entrepreneurs are shrewd political economists. They always try to establish strong relationships with governments at all levels and serve in formal political organizations such as the NPC (the National People’s Congress) and the CPPCC (the Chinese People’s Political Consultative Conference), or government-led industrial organizations like the ACFIC (All-China Federation of Industry and Commerce), to endear their business activities. Since political involvement typically improves firms’ social and political statuses, such firms are more motivated to become socially responsible to pay back their social power and to gain continuous political benefits in the future [69].

Second, political involvement may grant speculative firms shelter against potential penalties. Despite China’s rapid and great economic transformation, governmental intervention is likely to continue in the foreseeable future [72]. Thus, Chinese firms, especially privately owned firms, regard political involvement as an important and effective strategy to ensure smooth business operation [73]. Political involvement is able to bring speculative firms favorable regulatory conditions [62] as well as political resources [72,74], which allow the speculators to mask their self-serving activities and thus reduce the ensuing pressure or potential administrative punishment. Therefore, privately owned firms with political involvement and resources depend little on CSR as an insurance premium for seeking political protection.

Overall, speculative firms with political involvement reciprocate social power through responsible activities favored by governments, such as corporate philanthropy, to maintain continuous political benefits and mitigate negative influence from stakeholders, revealing that the effect of corporate speculation on CSR can be partly offset by that of political involvement.

Therefore, we can assume that:

Hypothesis 2.

The relationship between corporate speculation and CSR is less pronounced among firms with political involvement.

2.3.3. The Role of Regional Development

Firms are always influenced by the environment they are embedded in, which affects their behavior and the consequences of this behavior [75]. It is not difficult to infer that corporate speculation, the resulting CSR engagement and how they are linked is influenced by the environment firms operate in. The level of formal institutions in local regions is particularly influential. China is a vast country and is characterized by uneven regional development. Different provinces maintain different levels of social and economic development, and formal institutions also vary across the country. Generally speaking, provinces with a more developed economy enjoy more developed institutions such as open markets, fair competition, efficient administration and effective regulatory systems [76,77]. Uneven regional development provides a perfect setting to examine how environment influences firms’ behaviors and how CSR activities work differently in different environments.

Therefore, based on Hypotheses 1 and 2, we further examine the relationship between corporate speculation, CSR and informal institution, and how it varies in different regions with different formal institutional environments. Specifically, we believe that in provinces with a more developed economy and formal institutions, the instrumentality of CSR will be weakened, and so will the moderating role of informal institutions. A primary reason is that a more complete formal institutional system not only promotes local economic activity but also better controls unethical or unsustainable behaviors and restricts the effect of informal institutions. Therefore, speculative firms in relatively developed regions probably find CSR engagement is not as effective at masking speculation, meaning the link between corporate speculation and CSR is weakened. Additionally, political involvement may contribute little to mitigate speculation’s negative consequences or help obtain political resources; therefore, the role of informal institutions is also weakened.

Hence, we can assume that:

Hypothesis 3a.

The relationship between corporate speculation and CSR is less pronounced among developed regions.

Hypothesis 3b.

The moderating role of political involvement on the relationship between corporate speculation and CSR is less pronounced in more developed regions.

3. Research Methodology

3.1. Sample

The initial sample came from the 2014 Chinese national survey of privately owned enterprises, which was jointly conducted by the United Front Work Department of the CPC Central Committee, ACFIC and the Private Economy Research Institutes of China. The sampling method was a multi-stage random sampling, and the respondents were private enterprises chosen from 31 provinces, autonomous regions and municipalities in mainland China covering all 20 major national industrial sectors. The data set is widely used for studying the CSR of private enterprises in China [10,30,34,35,36,78,79].

The data on CSR, corporate speculation, entrepreneurs’ specific control variables and firm-specific control variables were obtained from this 2014 survey data set. In addition, we obtained the province-level GDP per capita data from the National Bureau of Statistics and the Marketization Index for China’s Provinces was obtained from Wang et al. [77]. The data were matched to each observation.

The initial data set included 6144 observations. Referring to previous studies [10,30,34,35], we selected our research sample in accordance with the following steps. First, we deleted the observations with unavailable or missing data on CSR and corporate speculative investment [10]. Second, we eliminated observations without available data on entrepreneurs’ attributes, firm attributes and other key control variables. Third, we removed the observations affiliated with the finance and real estate industries to reduce measurement error, as the speculative behavior defined in this study is their main business [34,35,36]. Fourth, to reduce the potential impact of outliers we winsorized the continuous control variables at the top and bottom 5% of the data [30,34]. We obtained a final sample of 2404 observations.

3.2. Empirical Models and Variables

We used Model Ι, shown below, to test how a private firm’s speculative behavior is associated with its CSR (Hypothesis 1) after controlling other variables including entrepreneurs’ attributes, firms’ attributes, industry dummies and institutional attributes:

$$\begin{array}{c}CSR={\alpha }_0+{\alpha }_{1}SPEINV+{\alpha }_{2}STATUS+{\alpha }_{3}EDU+{\alpha }_{4}AGE+{\alpha }_{5}FMOWN\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+{\alpha }_{6}FORST+{\alpha }_{7}SIZE+{\alpha }_{8}HIS+{\alpha }_{9}ROE+{\alpha }_{10}LEV\hfill \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+{\alpha }_{11}FORINV+{\alpha }_{12}GDP+{\alpha }_{13}MKT\hfill \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+Industry Dummies+\mathsf{\epsilon }\hfill \end{array}$$

(1)

(Model I).

In Model Ι the dependent variable is corporate social responsibility (CSR). Following extant literature on emerging economies that have taken corporate philanthropy (rather than other major components of Carroll’s pyramid of social responsibility) as a direct proxy of CSR [3,4,5,6,7,8], we measured CSR as the total amount of corporate philanthropic donations (multiplied by 1000) scaled by total sales in the year [30]. Moreover, we used alternative measures for the dependent variable in the robustness checks, including CSR_1 (a dummy variable coded as 1 if, in that year, a private firm made a philanthropic donation, or 0 if no donation was made) and CSR_2 (natural log of the total amount of corporate philanthropic donations). The independent variable is corporate speculation (SPEINV), a dummy variable coded as 1 if a private firm has invested in real estate, the stock market, the futures market or private lending, and otherwise 0. We also used the alternative measure SPEINV_1 (the total investment in real estate, the stock market, the futures market and private lending scaled by the year’s total sales) in robustness checks. Hypothesis 1 will be supported if the coefficient of SPEINV is positive and significant.

Referring to prior studies [30,34,35,36,57], we controlled the impact of three entrepreneur attributes including status (STATUS), education (EDU) and age (AGE) in Model Ι. STATUS is defined as an enterpriser’s self-perceived overall status, which is the average of the self-perceived economic, social and political status. EDU reflects the entrepreneur’s educational background. It is measured on a scale from 1 to 6, starting with primary school (represented as 1), then middle school, high school, college, university and postgraduate studies or above. AGE is the age of a private entrepreneur. Following the existing literature [10,38], we also controlled the seven firm-level attributes—ownership (FMOWN), governance structure (FORST), firm size (SIZE), firm history (HIS), return on equity (ROE), leveraged rate (LEV) and foreign investment (FORINV)— in Model Ι, reflecting firm corporate governance, size, history and basic financial background. FMOWN is measured as the shareholding equity ratio of the entrepreneur and their family. FORST considers if a private enterprise has a formal governance structure that includes shareholder meetings, a board of directors or a board of supervisors. We used 1 (3/3) for those with all the three structures; 0.667 (2/3) for those with two of the three; 0.333 (1/3) for those with one of the three; and 0 (0/3) for those with no such structures. SIZE is measured as the natural logarithm of the number of employees in that year. HIS is measured as the number of years since the firm‘s foundation. ROE is measured as the net income scaled by owner’s equity. LEV is measured as the total outstanding loans from banks, small financial institutions or peer-to-peer lending scaled by that years’ sales. FORINV is measured as the amount of foreign investment scaled by that year’s sales. Following previous studies, we also included GDP (GDP) and marketization index (MKT) to control regional heterogeneity [30,34,57]. GDP is measured as the GDP per capita (in 1000 yuan) at a provincial level. MKT is measured as the overall market development indices of Wang et al. [77] at a provincial level. In addition, we controlled for the industry’s fixed influence by adding dummy variables.

We used Model II below to examine the moderating effect of political ties (Hypothesis 2):

$$\begin{array}{c}CSR={\beta }_0+{\beta }_{1}SPEINV+{\beta }_{2}POLTIE+{\beta }_{3}OPPTIC\times POLTIE\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+{\beta }_{4}STATUS+{\beta }_{5}EDU+{\beta }_{6}AGE+{\beta }_{7}FMOWN\hfill \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+{\beta }_{8}FORST+{\beta }_{9}SIZE+{\beta }_{10}HIS+{\beta }_{11}ROE+{\beta }_{12}LEV\hfill \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+{\beta }_{13}FORINV+{\beta }_{14}GDP+{\beta }_{15}MKT\hfill \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+Industry Dummies+\mathsf{\epsilon }\hfill \end{array}$$

(2)

(Model II).

In the above model, we included corporate speculation (SPEINV), political involvement (POLITIE) and the interaction item (SPEINV $\times$ POLITIE), as well as entrepreneurs’ attributes, firm attributes, industry dummies and institutional attributes. POLITIE is a dummy variable coded as 1 if the entrepreneur or chairperson currently serves as a deputy in the NPC, or as a member of the CPPCC or ACFIC. Otherwise, it is 0. Hypothesis 2 will be supported if the coefficient of SPEINV × POLTIE is negative and significant. Other control variables are the same as those in Model Ι.

4. Empirical Results

4.1. Descriptive Statistics

Table 1. Descriptive statistical analysis of major variables.

Variables	Obs	Mean	Std. Dev.	Min	Q1	Median	Q3	Max
CSR	2404	0.109	0.189	0	0	0	0.148	0.740
CSR_1	2404	0.478	0.5	0	0	0	1	1
SPEINV	2404	0.124	0.33	0	0	0	0	1
SPEINV_1	2404	0.013	0.067	0	0	0	0	0.769
POLTIE	2404	0.572	0.495	0	0	1	1	1
STATUS	2404	5.336	1.78	1	4	5.333	6.667	10
EDU	2404	4.052	1.098	1	3	4	5	6
AGE	2404	46.262	8.876	18	40	46	52	79
FMOWN	2404	80.186	28.987	0	60	99	100	100
FORST	2404	0.429	0.348	0	0	0.333	0.667	1
SIZE	2404	4	1.73	0.693	2.708	4.06	5.193	10.653
HIS	2404	10.324	6.066	0	5	10	14	36
ROE	2404	1.366	3.07	−0.15	0.034	0.183	1	12.455
LEV	2404	0.375	0.736	0	0	0.067	0.329	2.935
FORINV	2404	0.003	0.01	0	0	0	0	0.044
GDP	2404	55.748	21.211	22.86	36.62	56.18	68.33	97.61
MKT	2404	7.353	1.759	−0.3	6.2	7.17	8.89	9.88

presents the descriptive statistics of the key variables. The mean (median) values of CSR and CSR_1 are 0.109 (0) and 0.478 (0), revealing that private firms’ average philanthropic donation-to-sales ratios are 0.0109% and 47.8% of the sample firms have made charitable donations. The mean (median) values of SPEINV and SPEINV_1 are 0.124 (0) and 0.013 (0), respectively, revealing that 12.4% of the sample firms engage in speculative behavior, and the average investment is about 1.3% of the total sales of private firms. The mean (median) value of POLTIE is 0.572 (1), revealing that more than half of Chinese private firms have some degree of connections with governments and official organizations.

Table 1 also confirms that the control variables show similar statistical characteristics with those in the current literature [10,30,33]. The average self-perceived status of the sampled Chinese private firms is moderate (mean STATUS = 5.336), with the highest being 10 and the lowest being 1. The average education level of the sampled entrepreneurs is a college degree (mean EDU = 4.052). The average age of entrepreneurs (AGE) is 46.262 years old, with the oldest being 79 and the youngest being 18. The average shareholder equity ratio of private entrepreneurs is 80.186% (mean FMOWN = 80.186), with the highest being 100% while the lowest is 0%. The mean (median) value of FORST is 0.429 (0.333), meaning most of the sampled private firms do not have a well-formed governance structure. The average number of employees is 55 (mean SIZE = 4), with the fewest being 6 and the most being 42,319. The average age of the sampled private firms (HIS) is 10.324 years, with the oldest firm being 36 years old and the newest being established for less than 1 year. The average return on equity is 136.6% (mean ROE = 1.366), with the highest being 1245.5% and the lowest being −15%. The average total loans of the sampled firms is 37.5% of the years’ sales (mean LEV = 0.375), with the highest being 293.5%. The average amount of foreign investment, as a percentage of that years’ sales, is 0.3% (mean FORINV = 0.003), with the highest being 4.4%. The average annual provincial GDP per capita is 55,748 yuan (mean GDP = 55.748), with the highest being 97,610 yuan. The average market development index (MKT) is 7.353, with the highest being 9.88.

4.2. Pearson Correlation Analysis

Table 2. Pearson correlation matrix.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)
(1) CSR	1.000
(2) CSR_1	0.604 *	1.000
(3) SPEINV	0.088 *	0.211 *	1.000
(4) POLTIE	0.244 *	0.360 *	0.103 *	1.000
(5) STATUS	0.156 *	0.253 *	0.113 *	0.338 *	1.000
(6) EDU	0.034	0.102 *	0.037	0.176 *	0.141 *	1.000
(7) AGE	0.096 *	0.144 *	0.069 *	0.186 *	0.191 *	−0.137 *	1.000
(8) FMOWN	−0.053 *	−0.039	0.014	−0.048 *	−0.071 *	−0.107 *	−0.054 *	1.000
(9) FORST	0.131 *	0.227 *	0.076 *	0.147 *	0.162 *	0.182 *	0.131 *	−0.212 *	1.000
(10) SIZE	0.254 *	0.492 *	0.217 *	0.444 *	0.376 *	0.235 *	0.249 *	−0.150 *	0.352 *	1.000
(11) HIS	0.174 *	0.262 *	0.106 *	0.321 *	0.268 *	0.031	0.394 *	−0.005	0.138 *	0.372 *	1.000
(12) ROE	0.078 *	0.205 *	0.161 *	0.142 *	0.136 *	0.116 *	0.057 *	0.014	0.146 *	0.279 *	0.123 *	1.000
(13) LEV	−0.073 *	−0.092 *	−0.036	−0.073 *	−0.101 *	−0.054 *	−0.030	0.001	−0.009	−0.115 *	−0.038	−0.042 *	1.000
(14) FORINV	−0.004	0.087 *	0.229 *	0.072 *	0.043 *	0.047 *	0.000	0.018	−0.008	0.088 *	0.075 *	0.037	−0.042 *	1.000
(15) GDP	−0.006	0.012	0.022	0.062 *	0.049 *	0.133 *	0.056 *	−0.004	−0.014	0.073 *	0.178 *	0.012	−0.087 *	0.066 *	1.000
(16) MKT	0.020	0.062 *	0.055 *	0.025	0.056 *	0.057 *	0.037	−0.014	−0.000	0.120 *	0.158 *	0.035	−0.061 *	0.048 *	0.794 *

Note: numbers with * indicate a significance level at 5% or better (two-ailed).

presents the Pearson correlations of the key variables. The results showed that our dependent variables, CSR and CSR_1, are significantly positively correlated with SPEINV, revealing a positive association between corporate speculation and CSR. In addition, a significantly positive association between POLTIE and CSR (CSR_1) was found. As for the control variables, the CSR displayed a significant association with STATUS, AGE, FMOWN, FORST, SIZE, HIS, ROE and LEV, revealing a need to control the influence of these variables when conducting further regression analyses. Moreover, the coefficient of each pairwise correlation among independent and control variables is less than 0.5, meaning there is no serious multicollinearity problem when conducting further empirical tests.

4.3. Multivariate Analysis

(1) Multivariate test of Hypothesis 1

To test our hypotheses, we had to consider the particular nature and distribution of the dependent variable of CSR, which is a continuous variable bound between 0 and 0.74. Moreover, our sample contained a high proportion of 0 values (52.2%), so we adopted a Tobit regression methodology and set 0 as a lower bound [80]. This methodological approach has been used frequently in similar contexts to deal with the censored nature and peculiar distribution of a dependent variable [30,37].

Table 3. Regression Results of CSR on corporate speculation and other determinants.

Variables	(1)	(2)
SPEINV		0.063 ***
		(2.98)
STATUS	0.014 ***	0.013 ***
	(3.02)	(2.95)
EDU	−0.005	−0.004
	(−0.65)	(−0.62)
AGE	−0.001	−0.001
	(−0.57)	(−0.60)
FMOWN	0.000	0.000
	(0.30)	(0.18)
FORST	0.056 **	0.056 **
	(2.54)	(2.52)
SIZE	0.064 ***	0.062 ***
	(11.13)	(10.75)
HIS	0.007 ***	0.007 ***
	(4.78)	(4.76)
ROE	0.002	0.002
	(1.06)	(0.73)
LEV	−0.032 ***	−0.032 ***
	(−2.88)	(−2.87)
FORINV	0.323	−0.155
	(0.49)	(−0.23)
GDP	−0.001 *	−0.001 *
	(−1.88)	(−1.76)
MKT	0.007	0.006
	(1.00)	(0.82)
Constant	−0.362 ***	−0.351 ***
	(−4.96)	(−4.82)
Industry dummy	Yes	Yes
Observations left-censored	2404	2404
Observations	1254	1254
Pseudo R2	0.2063	0.2101
Chi-square	482.13	491.02

Note: ***, ** and * represent a significance level at 1%, 5% and 10% or better, and the value in the brackets is the t-values.

presents the multiple regression analyses conducted on the sample firms to test for Hypothesis 1. Column (1) shows the Tobit regression result of CSR on the control variables. Column (2) shows Tobit regression results of CSR on SPEINV and the control variables.

As shown in column (2) of Table 3, the coefficient of SPEINV (β = 0.063, t = 2.98) is positive and statistically significant at the 0.01 level. Hypothesis 1 is strongly supported. Moreover, the results revealed two findings. Firstly, corporate speculation is significantly and positively associated with CSR engagement, indicating private Chinese firms may use philanthropy as a kind of ‘profit-sharing’ tool to mask their speculative activities and then exchange reputation to alleviate negative assessment. Second, speculative firms are motivated to participate in CSR, which echoes the risk-managing explanation of motives for CSR engagement in extant literatures [18].

With regard to the control variables, column (2) of Table 3 shows that the coefficients of FORST (β = 0.056, t = 2.52), SIZE (β = 0.062, t = 10.75) and HIS (β = 0.007, t = 4.76) are positive and significant at the 0.05, 0.01 and 0.01 levels respectively, which echoes previous studies [38,63]. Moreover, the coefficients of LEV (β = −0.032, t = −2.87) and GDP (β = −0.001, t = −1.76) are negative and significant at the 0.01 and 0.1 levels. These results oppose the findings of Du’s study [30]. The impact of the other control variables is not significant, as is shown in Table 3. These results of both column (1) and (2) are consistent, which reinforces the importance of controlling firm and regional attributes when testing how a private firm’s speculative behavior is associated with CSR.

(2) Multivariate test of Hypothesis 2

We conducted a step-by-step regression analysis of CSR on POLTIE, SPEINV and the control variables to test Hypothesis 2.

Table 4. Regression results of CSR on corporate speculation, political ties and other determinants.

	(1)	(2)	(3)
SPEINV		0.065 ***	0.183 ***
		(3.13)	(4.85)
POLTIE	0.142 ***	0.142 ***	0.166 ***
	(8.30)	(8.35)	(9.08)
SPEINV × POLTIE			−0.163 ***
			(−3.73)
STATUS	0.007	0.007	0.006
	(1.58)	(1.50)	(1.44)
EDU	−0.010	−0.009	−0.010
	(−1.35)	(−1.31)	(−1.38)
AGE	−0.001	−0.001	−0.001
	(−0.85)	(−0.88)	(−0.95)
FMOWN	0.000	−0.000	−0.000
	(0.09)	(−0.03)	(−0.04)
FORST	0.058 ***	0.057 ***	0.059 ***
	(2.65)	(2.63)	(2.71)
SIZE	0.052 ***	0.050 ***	0.049 ***
	(9.02)	(8.62)	(8.53)
HIS	0.005 ***	0.005 ***	0.005 ***
	(3.59)	(3.57)	(3.61)
ROE	0.002	0.001	0.002
	(1.00)	(0.65)	(0.84)
LEV	−0.032 ***	−0.032 ***	−0.031 ***
	(−2.89)	(−2.88)	(−2.84)
FORINV	0.224	−0.272	−0.377
	(0.35)	(−0.41)	(−0.57)
GDP	−0.001 **	−0.001 **	−0.001 **
	(−2.29)	(−2.16)	(−2.27)
MKT	0.011	0.010	0.010
	(1.63)	(1.44)	(1.47)
Constant	−0.350 ***	−0.340 ***	−0.347 ***
	(−4.84)	(−4.70)	(−4.79)
Industry dummy	YES	YES	YES
Observations left-censored	2404	2404	2404
Observations	1254	1254	1254
Pseudo R2	0.2361	0.2403	0.2463
Chi-square	551.97	561.76	575.65

Note: ***, and ** represent a significance level at 1%, 5% or better, and the value in the brackets is the t-values.

, column (1) shows the Tobit regression results of CSR on POLTIE and the control variables; column (2) adds the independent variables, SPEINV, into the regression model; and column (3) incorporates the interaction between SPEINV and POLTIE into the regression model.

The results in Table 4, column (1) show that the coefficient of POLTIE (β = 0.142, t = 8.30) is positive and significant at the 0.01 level. This reveals that private firms with political involvement are more likely to have high social consideration, which motivates them to engage in CSR. This finding echoes the previous 2015 study by Li et al. [37]. In column (2), the Tobit regression results show that the coefficients of SPEINV (β = 0.065, t = 3.31) and POLTIE (β = 0.142, t = 8.35) are positive and significant at the 0.01 level, which is consistent with the findings in column (1) in Table 3 and Table 4, respectively.

Furthermore, from the regression results in column (3) of Table 4, we see the coefficient of the interaction term (SPEINV × POLTIE) (β = −0.163, t = −3.73) is negative and significant at the 0.01 level. This is consistent with our expectation that a firm’s political involvement influences the relationship between speculative behavior and CSR. The positive association between corporate speculation and CSR is less pronounced in private firms with political involvement. This finding is consistent with the opinions of current literature that political involvement provides special access to resources and legitimacy and thus reduces the pressure from society and government to act ethically, as reported by Du in 2015 [30], and Li et al. in 2015 [37]. Additionally, the coefficient of SPEINV (β = 0.183, t = 4.85) is significantly positive, consistent with findings in columns (1) and (2). This largely supports Hypothesis 2.

(3) Multivariate test of Hypothesis 3

The regression results in both Table 3 and Table 4 show the GDP coefficient is consistently negative and significant. This suggests that the regional development level is likely to influence the associations among corporate speculation, political involvement and CSR. To address the concern that regional development level may influence our results, we conducted another regression analysis, Model II, of CSR on SPEINV, POLTIE and control variables, which incorporated the level of regional development. A region where GDP ≥ 90 (in 1000 yuan) was regarded as developed and a region where GDP < 90 (in 1000 yuan) was regarded as developing. The sample was divided into two sub-groups. The developed region sub-group contained 199 observations, and the developing regions sub-group contained 2205 observations.

Table 5. Regression results of CSR on corporate speculation, political ties and other determinants (developed regions vs developing regions).

	(1)	(2)
Variables	Developed Regions	Developing Regions
SPEINV	0.266	0.183 ***
	(1.62)	(4.76)
POLTIE	0.271 ***	0.165***
	(3.78)	(8.70)
SPEINV × POLTIE	−0.246	−0.164 ***
	(−1.40)	(−3.66)
STATUS	0.015	0.006
	(1.00)	(1.27)
EDU	−0.023	−0.010
	(−0.99)	(−1.32)
AGE	−0.005	−0.001
	(−1.37)	(−0.61)
FMOWN	−0.001	0.000
	(−0.68)	(0.02)
FORST	0.204 ***	0.049 **
	(2.92)	(2.17)
SIZE	0.046 **	0.049 ***
	(2.48)	(8.06)
HIS	0.015 ***	0.004 ***
	(3.19)	(2.79)
ROE	−0.016 **	0.003
	(−2.26)	(1.39)
LEV	−0.019	−0.031 ***
	(−0.46)	(−2.75)
FORINV	5.198 ***	−1.155 *
	(2.72)	(−1.65)
MKT	0.079	0.000
	(0.82)	(0.08)
Constant	−1.017	−0.338 ***
	(−1.09)	(−4.57)
Industry dummy	Yes	Yes
Observations	199	2205
Left-censored observations	110	1144
Pseudo R2	0.5554	0.2426
Chi-square	105.03	520.85

Note: ***, ** and * represent a significance level at 1%, 5% and 10% or better, and the value in the brackets is the t-values.

reveals a very interesting result. As we expected, the influence of SPEINV on CSR varies greatly across regions with different levels of economic development, and so does the moderating role of POLITIE. Specifically, column (1) presents that the coefficient of SPEINV (β = 0.266, t = 2.62) and the interaction term (β = −0.246, t = −1.40) is insignificant in the developed sample, while in column (2) the results reveal that speculative private firms in developing regions are more likely to participate in CSR. Moreover, the association between corporate speculation and CSR is weaker among private firms with political resources in developing regions. This result reveals a remarkable feature of transitional environments, and further confirms that regional and economic development can restrict the instrumentality of CSR engagement. These results provide further explanations and support for Hypotheses 1 and 2.

4.4. Robustness Checks

We also conducted several robustness checks. First, we used alternative proxies for the dependent variables. Following previous research [10,37], we used the Logit model to run a multiple regression test with CSR_1 as the dependent variable, and the Tobit model with CSR_2 as the dependent variable. Second, we referred to Gao’s research [70] by using SPEINV_1 as an alternative proxy for the independent variable. The overall results are summarized in

Table 6. Robustness checks: using alternative dependent variables.

	(1)	(2)
	CSR_1	CSR_2
	Logit	Tobit
SPEINV	1.440 ***	4.626 ***
	(5.39)	(5.71)
POLTIE	0.955 ***	3.983***
	(8.16)	(10.14)
SPEINV × POLTIE	−1.075 ***	−4.706 ***
	(−3.28)	(−4.99)
STATUS	0.038	0.149
	(1.23)	(1.53)
EDU	−0.061	−0.197
	(−1.24)	(−1.28)
AGE	−0.008	−0.021
	(−1.25)	(−1.04)
FMOWN	0.003	0.003
	(1.53)	(0.49)
FORST	0.532 ***	1.875 ***
	(3.47)	(3.98)
SIZE	0.555 ***	1.960 ***
	(12.68)	(15.64)
HIS	0.030 ***	0.107 ***
	(3.13)	(3.62)
ROE	0.059 ***	0.145 ***
	(3.13)	(3.02)
LEV	−0.146 **	−1.029 ***
	(−2.00)	(−4.25)
FORINV	4.773	−3.708
	(1.01)	(−0.26)
GDP	−0.012 ***	−0.027 **
	(−2.93)	(−2.22)
MKT	0.110 **	0.272 *
	(2.33)	(1.83)
Constant	−3.315 ***	−10.521 ***
	(−6.61)	(−6.72)
Industry dummy	YES	YES
Observations	2404	2404
Left-censored observations	--	1254
Pseudo R2	0.2569	0.1047
Chi-square	855.15	1050.00

Note: ***, ** and * represent a significance level at 1%, 5% and 10% or better, and the value in the brackets is the t-values.

and

Table 7. Robustness checks: using alternative independent variables.

	(1)	(2)	(3)
SPEINV_1	0.476 **	4.973 *	0.472 **
	(2.38)	(1.79)	(2.36)
POLTIE	0.147 ***	0.275 ***	0.146 ***
	(8.47)	(3.86)	(8.08)
POLTIE × SPEINV_1	−0.432 *	−5.016	−0.413 *
	(−1.89)	(−1.66)	(−1.79)
STATUS	0.007	0.014	0.007
	(1.58)	(0.88)	(1.42)
EDU	−0.010	−0.025	−0.010
	(−1.37)	(−1.06)	(−1.32)
AGE	−0.001	−0.004	−0.001
	(−0.92)	(−1.29)	(−0.57)
FMOWN	0.000	−0.001	0.000
	(0.06)	(−0.61)	(0.10)
FORST	0.060 ***	0.220 ***	0.049 **
	(2.72)	(3.14)	(2.16)
SIZE	0.052 ***	0.044 **	0.052 ***
	(9.04)	(2.40)	(8.59)
HIS	0.005 ***	0.016 ***	0.004 ***
	(3.62)	(3.30)	(2.79)
ROE	0.002	−0.016 **	0.003
	(0.97)	(−2.25)	(1.49)
LEV	−0.032 ***	−0.019	−0.032 ***
	(−2.88)	(−0.45)	(−2.81)
FORINV	0.078	5.635 ***	−0.676
	(0.12)	(3.04)	(−0.98)
GDP	−0.001 **	—	—
	(−2.19)
MKT	0.011	0.081	0.001
	(1.53)	(0.84)	(0.28)
Constant	−0.352 ***	−1.051	−0.344 ***
	(−4.86)	(−1.13)	(−4.65)
Industry dummy	YES	YES	YES
Observations	2404	199	2205
Left-censored observations	1254	110	1144
Pseudo R2	0.2386	0.56	0.2347
Chi-square	557.65	105.59	503.75

Note: ***, ** and * represent a significance level at 1%, 5% and 10% or better, and the value in the brackets is the t-values.

.

Table 6 presents the results of robustness checks using alternative dependent variables. Columns (1) and (2) show that SPEINV has a significantly positive association with CSR_1 and CSR_2. The interaction coefficients are negative and significant at the 0.01 level. These results reinforce the primary findings and largely support our hypotheses.

Table 7 shows the results of the robustness checks using the alternative independent variable in the full sample and the developed and developing sub-samples. Column (1) shows that SPEINV_1 is still significantly positively associated with CSR and the coefficient of interaction between SPEINV_1 and POLTIE is negative and significant at the 0.1 level, echoing the findings in column (2) of Table 3.

Furthermore, column (2) of Table 7 shows that the coefficient of SPEINV_1 is significantly positive, which is inconsistent with the results in column (1) of Table 5. However, as expected, the coefficient of the interaction term is insignificant in the developed sample. Furthermore, column (3) of Table 7 shows that, in the developing sample, the coefficient of SPEINV_1 is positive and significant and the coefficient on interaction item is negative and significant. This is consistent with the findings in column (2) of Table 5. Despite some slight differences, the results in Table 7 suggest the main findings are not qualitatively changed by using an alternative independent variable.

There is a potential endogenous problem, as the relative timing of firms’ corporate speculative and CSR activities, which both form part of the firms’ revenue allocation, are not identified in the survey data. Following previous research [10,30,79], we used Newey’s two-step estimation to address this concern. We selected EFIN, measured as environmental fines scaled by that year’s sales, as an instrumental variable to capture exogenous variation in the analysis.

Table 8. Robustness checks: the second-stage of Tobit regression.

	(1)	(2)
SPEINV	0.565 ***	1.001 ***
	(2.75)	(3.48)
POLTIE		0.238 ***
		(5.29)
SPEINV × POLTIE		−0.753 **
		(−2.20)
STATUS	0.011 **	0.003
	(2.18)	(0.65)
EDU	−0.002	−0.010
	(−0.30)	(−1.26)
AGE	−0.001	−0.001
	(−0.83)	(−1.27)
FMOWN	−0.000	−0.000
	(−0.45)	(−0.55)
FORST	0.053 **	0.064 ***
	(2.16)	(2.61)
SIZE	0.046 ***	0.034 ***
	(5.25)	(3.88)
HIS	0.006 ***	0.005 ***
	(4.22)	(3.28)
ROE	−0.004	−0.001
	(−1.17)	(−0.28)
LEV	−0.033 ***	−0.031 **
	(−2.71)	(−2.54)
FORINV	−3.572 **	−3.458 **
	(−2.27)	(−2.22)
GDP	−0.000	−0.001
	(−0.73)	(−1.55)
MKT	−0.002	0.004
	(−0.29)	(0.48)
Constant	−0.285 ***	−0.301 ***
	(−3.39)	(−3.58)
Industry dummy	YES	Yes
Observations	2404	2404
Left-censored observations	1254	1254
Wald test of exogeneity	7.36 **	10.24 **

Note: ***, and ** represent a significance level at 1%, 5% or better, and the value in the brackets is the t-values.

presents the results of the second-stage Tobit regressions. Columns (1) and (2) show that the coefficients of SPEINV are positive and significant at the 0.01 level, which echoes the findings in column (2) of Table 3. These results largely support Hypothesis 1. Column (2) also shows that the coefficient of interaction term is negative and significant at the 0.01 level, strongly supporting Hypothesis 2. Moreover, the coefficients of EFIN in both the 2SLS and Tobit regression are significantly positive, suggesting there is no weak instrumental variable. Therefore, the results indicate that political involvement reduces the positive association between corporate speculation and CSR even after controlling the endogeneity of SPEINV.

5. Discussion and Conclusions

To determine why rational, profit-oriented firms generously engage in CSR, we have explored the real motives of speculative firms for CSR engagement and the hidden causality behind it. Using a sample of 2404 privately owned firms in China, the empirical results show that corporate speculation is significantly and positively related to CSR engagement. This reveals that privately owned firms instrumentally use CSR to manage potential risks from their speculative activities, adding an explanation to the instrumentality of CSR in extant studies [19,21]. We also provided systematic evidence to show the substitutive effects of political involvement and corporate speculation on CSR engagement. Moreover, our study has further demonstrated that the speculation–CSR link and the influence of political involvement are more pronounced in relatively less-developed regions, and that the related link becomes less significant as economic growth and regional prosperity increase, revealing that the development of formal institutions may restrict the instrumentality of CSR and the effect of political involvement.

5.1. Theoretical Implications

Our study contributes to the body of literature on the motives of CSR engagement. It appears that this is the first study to investigate the hidden connection between speculation and CSR engagement in an emerging economy, which thus provides a more comprehensive view of the instrumentality of CSR as compared with extant studies [4,5,19,21]. Our results have confirmed that speculative firms are instrumentally motivated to engage in CSR due to hidden non-monetary benefits in which CSR engagement does appear to mitigate risks to reputation-building and acquire insurance-like protection in China’s emerging economy [18]. More specifically, privately owned firms try to utilize CSR as a tool to mask their speculative activities by building their reputations and obtaining ’leniency insurance’ against potential penalties. However, the insurance-like protection arising from CSR engagement can be substituted for direct political protection from officers and regulators, revealing that informal institutions can guard against potential sanctions without addition premiums. Furthermore, the above mechanisms vary across regions with different levels of social and economic development, thereby suggesting that the instrumental motives for CSR might weaken gradually with economic development.

5.2. Managerial Implications

Disclosing the association between corporate speculation and CSR engagement has many implications for CSR policies and practices in China as well as other emerging economies.

First, our study shows that in emerging China, the cost of corporate speculation is relatively low, and firms pursue social engagement primarily for the purposes of masking their speculative behaviors, using it as insurance-like protection against potential sanctions. We note that such CSR engagement is, on its face, honorable, which means that we should be mindful when honoring a firm for its philanthropy and other ostensibly responsible activities. We should set up a more developed evaluation system to differentiate substantive CSR and purely instrumental CSR, and we should praise and reward firms for their engagement in the former.

Second, our study shows that the speculation–CSR relationship is shaped or disrupted by informal institutions. Political involvement, as a critical informal institution in emerging economies, sometimes serves as a shelter for corporate speculation. Thus, when considering long-term social development, we must rectify firms’ understandings of CSR. A top priority is to create a level playing field for all firms by restricting governmental intervention into business, and improving market-oriented mechanisms to enhance formal institutions.

Third, our findings on CSR in transitioning China might have implications for other emerging countries that are undertaking social, economic and institutional transformations. It must be noted that every country has been influenced by its unique cultural and social environment. It is critical to explore how such environmental factors work from a dynamic perspective, which surely provides guidance for wise policymaking and regulations. When exploring the motives of CSR, we must take national conditions into consideration, such as the level of economic and market development and the role of governments.

5.3. Limitations and Future Research

Although it has provided unique and important insight into CSR in emerging economies, our study was subject to certain limitations that are expected to be addressed in future research. First, some of the results may not be widely generalized. In countries where markets—rather than governments—drive resource flows, and where formal institutions—rather than informal institutions—play a dominant role in business administration, the situation might be very different. Future studies could examine whether corporate speculation remains positively associated with CSR engagement in other emerging economies with relatively developed market mechanisms and less intervention from informal institutions. Second, it is reasonable to use philanthropic donations as a proxy of CSR at China’s current stage of development, but this biased measurement may not catch all features of CSR and might not be suitable for all developing countries. Future studies could present a more comprehensive picture of CSR by incorporating other components of Carroll’s pyramid of social responsibility [2] into the analysis. Finally, due to the limitations of the survey data, we have conducted an empirical analysis based on a one-year data set. Thus, the results only capture the recent characteristics and motives of CSR, while they are not able to present the dynamics of firm behaviors when private firms continuously ‘fit’ themselves into the external environment. Although it would be useful to analyze data from a variety of geographical regions to show the dynamics, longitudinal data would be more convincing in presenting related dynamics. Future studies could collect panel data to re-examine and enrich the findings of our current study.