The marital demography in China

For thousands of years, given the effect of Confucian culture on the Chinese society, the structure within the family was extremely stable (Fei, Fei, Hamilton, & Zheng, Reference Fei, Fei, Hamilton and Zheng1992, pp. 87–93) and divorce was rare in China. Since the founding of the People's Republic of China, women had been given the unprecedented status, known as ‘Women can hold up half of the sky’ (Du, Reference Du2016), and thus the divorce has become common in China. According to China's National Bureau of Statistics (www.stats.gov.cn), in the mid-1980s, the average divorce–marriage ratio (divorce rate) was about 6.82% (.68‰), but the highest ratio (rate) at the province level was 34.99% (3.64‰).

Along with China's deepening of the Reform and Opening-Up Policy, the weakened influence of Confucian culture on the Chinese society, and the impact of foreign culture, the mean (maximum) values of the divorce–marriage ratio and divorce rate in 1999 reached 16.90% (35.62%) and 1.11‰ (3.29‰).Footnote ¹ In 2017, the average divorce–marriage ratio (divorce rate) was 41.64% (3.14‰), along with the maximum value of 68.97% (5.20‰). Overall, the average divorce–marriage ratio (divorce rate) in most provinces of mainland China embodies the steady increased tendency. Figure 1a (1b) outlines the changing tendency of the divorce–marriage ratio (divorce rate) from the mid-1980s to 2017, revealing the marital demography in contemporary China.

Figure 1. The distribution characteristics of marital demography. (a) The tendency of the divorce–marriage ratio during the period of 1985–2017 (unit: %; the shaded part denotes the sample period), (b): The tendency of the divorce rate during the period of 1985–2017 (unit: ‰; the shaded part denotes the sample period), (c): China's provincial distribution of the divorce–marriage ratio in 2017 (unit: %), (d): China's provincial distribution of the divorce rate in 2017 (unit: ‰), (e): China's provincial distribution of the mean value of the divorce–marriage ratio during the sample period (2008–2017; unit: %), (f): China's provincial distribution of the mean value of the divorce rate during the sample period (2008–2017; unit: ‰).

Given the close relation between marriage (divorce) and collectivism (individualism) (Vandello & Cohen, Reference Vandello and Cohen1999), the continuous upward divorce–marriage ratio (divorce rate) in the past decades embodies the change in social culture. The Chinese society is always characterized with higher (lower) extent of collectivism (individualism) (Chen, Meindl, & Hunt, Reference Chen, Meindl and Hunt1997; Earley, Reference Earley1989), but this impression seems to be undergoing the challenge (Koch & Koch, Reference Koch and Koch2007). In this regard, the divorce–marriage ratio (divorce rate) reflects the dynamic and evolving attitude and the weakened responsibility towards the matrimony (family) (Du, Reference Du2016; Vandello & Cohen, Reference Vandello and Cohen1999).

Motivated by above settings, our study aims to investigate whether individualistic social culture – which is embedded in increasingly high divorce–marriage (divorce) ratio – affects greenwashing.

Environmental conservation and greenwashing in China

Greenwashing originated from the American hotel industry (Kenton, Reference Kenton2020), but the hotels in the world just enjoy the benefit of lower laundry costs and do not conduct sufficient eco-friendly activities. Inspired by the hotel industry, a variety of corporations in various industries claim that their products, services, and even the whole enterprise are environmentally friendly (Du, Reference Du2015). However, a great number of corporations are only talk the talk rather than walk the walk. Against this context, corporate greenwashing becomes increasingly prevalent today (Greenpeace USA, 2013; Parguel, Benoît-Moreau, & Larceneux, Reference Parguel, Benoît-Moreau and Larceneux2011).

Compared with western countries, greenwashing behavior is not inferior in China. Du et al. (Reference Du, Jian, Zeng and Du2014, Reference Du, Chang, Zeng, Du and Pei2016, Reference Du, Jian, Zeng and Chang2018) and Du (Reference Du2015) reveal that environmental performance of Chinese listed firms is far from optimistic. Specifically, Chinese listed firms tend to score about 3.53 out of 95 (the full score) on average and the highest score is about 44 – less than half of 95. In this regard, it is a huge irony that a great number of Chinese listed firms have declared their eco-friendliness. More likely, hundreds of enterprises conduct greenwashing to cover their environmental weaknesses or unfriendliness.

Greenwashing has penetrated into food, medical treatment, and public welfare products. Du (Reference Du2015) and the South Weekend (www.infzm.com; China's most influential unofficial newspaper) summarize Chinese-style greenwashing: (1) a naked liar (open fraudulence); (2) deliberately concealing facts; (3) double-standard environmental behavior; (4) empty promises; (5) environmental policy interference; (6) environmental impression management; and (7) negative environmental externality.

As environmental pollution becomes increasingly prominent and greenwashing intensifies, some unofficial organizations and the media step forward bravely to disclose the lists of firms that conduct greenwashing behavior. For example, with the support of careful investigations, the South Weekend has published the yearly list of firms with greenwashing since 2012 (Du, Reference Du2015). It is amazing that some very famous Chinese enterprises (even multinational enterprises) appear more than once in these lists (e.g., China Sinopec, 600028.SH). In recent year, due to the continuous attention to carbon emission and the greenhouse effect, China Sinopec – one of the largest oil companies conducted greenwashing through rebranding itself as the environmental pioneer. Specifically, China Sinopec renamed its services, repacked its products, and even advertised its green or environmental image during the processes of charitable donations and conveying the sympathy to the victims of natural disasters.

Prior literature on the marital demography

Family is the most basic component of the human society (Morgan & Morgan, Reference Morgan and Morgan1996), the matrimony affects family composition, and in turn family composition shapes social stability and social culture (Olson & DeFrain, Reference Olson and DeFrain2003, pp. 100–105; Sanders & Nee, Reference Sanders and Nee1996). In recent year, scholars in the fields of economics and management borrow the proxies for marital demography from the sociology (demography) to examine their effect on individual behavior and corporate decisions.

Prior studies have reached consistent conclusions that marriage (divorce) can improve (reduce) individual subjective well-being. Gove, Hughes, and Style (Reference Gove, Hughes and Style1983) find that the marital status is the most powerful predictor of the mental health. Gove and Shin (Reference Gove and Shin1989) validate that, compared to the married people, the psychological well-being of the divorced (widowed) is far poorer. Stutzer and Frey (Reference Stutzer and Frey2006) focus on the relationships between marriage and subjective well-being, validating that happier singles are more likely to opt for marriage. Shapiro and Keyes (Reference Shapiro and Keyes2008) find that married persons have a significant social well-being advantage, in addition to the positive link between marriage and psychological well-being.

Marital status also strengthens philanthropic giving and reduces the likelihood of crime. Bryant, Jeon-Slaughter, Kang, and Tax (Reference Bryant, Jeon-Slaughter, Kang and Tax2003) argue that the marital status is an important dimension of social capital, and thus married people are more connected with social networks than single, separated, and divorced people – who are less likely to be volunteers or make philanthropic giving. Einolf and Philbrick (Reference Einolf and Philbrick2014) and Barnes and Beaver (Reference Barnes and Beaver2012) verify that marriage has a stronger positive (negative) effect on giving (crime).

Besides, a branch of research examines the influence of marital status on the attitudes towards the risk and managerial behavior. Hartog, Ferrer-i-Carbonell, and Jonker (Reference Hartog, Ferrer-i-Carbonell and Jonker2002) argue that the married people are more risk averse than their counterparts because the marriage contract increases the costs of breaking up the relationship. Roussanov and Savor (Reference Roussanov and Savor2014) find that marital status affects individual preferences towards risk, and firms with single (married) CEOs always have higher (lower) stock volatility and exhibit more (less) aggressive investment. Christiansen, Joensen, and Rangvid (Reference Christiansen, Joensen and Rangvid2010) conclude that marriage is a financial risk-reducer for men and financial risk-increaser for women. Nicolosi and Yore (Reference Nicolosi and Yore2015) validate that mergers and major capital expenditures are significantly related with divorces (marriages), and especially, the link between a CEO's marital status and the preference for option-based compensation. Neyland (Reference Neyland2016) verifies that divorce has significant effect on firms with cash-poor CEOs. Hilary, Huang, and Xu (Reference Hilary, Huang and Xu2017) find that firms with single CEOs display a higher degree of earnings management than firms with married CEOs.

Lastly, another focus in extant studies addresses the impact of marital status on social culture. Vandello and Cohen (Reference Vandello and Cohen1999) find that states with higher marriage–divorce ratios are more collectivist in the USA. Toth and Kemmelmeier (Reference Toth and Kemmelmeier2009) find that divorce is more likely to link with individualist societies. As a matter of fact, the divorce rate is consistently associated with the individualistic cultural tendency within a region or society, and highly individualist regions (states, societies) always exhibit higher divorce rates (Dion & Dion, Reference Dion and Dion1996; Lester, Reference Lester1995).

Prior literature on corporate greenwashing

Greenwashing means a variety of different misleading communications that aim to form overly positive beliefs among stakeholders about a company's environmental practices (Torelli, Balluchi, & Lazzini, Reference Torelli, Balluchi and Lazzini2020). Thus, greenwashing is a firm's unsubstantiated claims, aiming at inducing stakeholders to believe that its products and services are energy-saving orientation and the whole enterprise is eco-friendly (Kenton, Reference Kenton2020). However, these claims could at most be partially believed, along with exaggerating the work in environmental protection and concealing environmental misconducts to the greatest extent (Du, Reference Du2015). In response, a branch of prior literature addresses the ways of identifying a firm's greenwashing.

Unfulfilled environmental commitment is closely linked with greenwashing (Lee & Hageman, Reference Lee and Hageman2018). Parguel et al. (Reference Parguel, Benoît-Moreau and Larceneux2011) argue that unfulfilled CSR claims and eco-friendly commitment motivate greenwashing, and further independent sustainability ratings can deter greenwashing and encourage firms to engage in eco-friendly activities. Laufer (Reference Laufer2003) argues that fair social reporting is a channel of identifying greenwashing because it reflects whether a firm has abided by environmental laws. In addition, no independent verification induces greenwashing. There is no statutory requirement that independent third party should verify environmental information, so large and leading-edge firms have lower likelihood of involving in greenwashing (Berrone, Fosfuri, & Gelabert, Reference Berrone, Fosfuri and Gelabert2017; Ramus & Montiel, Reference Ramus and Montiel2005). Du (Reference Du2015) adopts the published lists of firms in the South Weekend to proxy for greenwashing. Lyon and Maxwell (Reference Lyon and Maxwell2011) verify that the threat of environmental audit drives a firm's greenwashing. Lyon and Montgomery (Reference Lyon and Montgomery2013) find that social media can mitigate corporate greenwashing. Delmas and Burbano (Reference Delmas and Burbano2011) argue that managers, NGOs, and policymakers have the obligations to mitigate greenwashing. Siano, Vollero, Conte, and Amabile (Reference Siano, Vollero, Conte and Amabile2017) extend the greenwashing taxonomy by identifying a new type of irresponsible behavior, namely ‘deceptive manipulation’, which strengthens the ‘communicative constitution of organizations’ perspective. By combining the signaling theory with the legitimacy theory, Seele and Gatti (Reference Seele and Gatti2017) argue that greenwashing epistemologically depends on an external accusation, which is viewed as a distortion factor altering the signal reliability of green messages. And then, Seele and Gatti (Reference Seele and Gatti2017) propose a revised definition of greenwashing as cocreation of an external accusation toward an organization with regard to presenting a misleading green message.

Greenwashing is conducted to capitalize on increasing demands for eco-friendly image, products, and services from the society, regardless of whether the ways or channels of achieving these objectives are recyclable, free of environmental pollutions, or nonwasteful (Kenton, Reference Kenton2020). In other words, the only purpose of greenwashing is to deceive the society into believing that the firm and the green label are inseparable, and then grab the huge profits at the expense of the whole society. Thus, another branch of prior literature has investigated the economic consequences of corporate greenwashing.

Greenwashing leads to negative market reactions. Du (Reference Du2015) validates that greenwashing results in significantly negative cumulative abnormal returns around the exposure of greenwashing. Moreover, greenwashing elicits the skepticism from external auditors and customers. Du et al. (Reference Du, Jian, Zeng and Chang2018) find that greenwashing is significantly positively associated with a firm's receiving modified audit opinion. Chen & Chang (2013) verify that greenwashing brings out green consumer confusion and increases green perceived risk. Rahman, Park, and Chi (Reference Rahman, Park and Chi2015), Bloch and Banjeree (Reference Bloch and Banjeree2001) find that the ulterior motive in greenwashing brings out consumer skepticism about environmental claims. Lastly, Bowen (Reference Bowen2014) finds the reduction in environmentalism after corporate greenwashing. Torelli et al. (Reference Torelli, Balluchi and Lazzini2020) find that different levels of greenwashing have a significantly different influence on stakeholders’ perceptions of corporate environmental responsibility and stakeholders’ reactions to environmental scandals.

To sum up, prior literature has investigated the determinants and consequences of greenwashing, but it remains unknown about whether individualistic social culture – which is embedded into and evolves along with marital demography (the divorce–marriage ratio) – affects a firm's greenwashing.

The marital demography and greenwashing (Hypothesis 1)

Confucian culture in China can be classified as being collectivistic (Du, Reference Du2016), which embodies as stable material status and extremely low divorce rate or divorce–marriage ratio. However, as the implementation of family planning policy and the impact of western culture on contemporary China, the big family in the traditional sense no longer exists, young people under the one-child policy have become more and more selfish, and thus individualistic culture is on the rise and collectivistic culture is declining (Evans, Reference Evans and Kipnis2012; Kipnis, Reference Kipnis2012). An obvious example is the increasing divorce rate (see Figure 1). In this regard, the divorce–marriage ratio can serve as an appropriate proxy for individualism in China.

The divorce rate is a product of social and cultural transformation from the collectivism to the individualism (Inglehart, Reference Inglehart1997). Collectivism refers to a social pattern of closely linked individuals who view themselves as interdependent members within a collective (e.g., family; Vandello & Cohen, Reference Vandello and Cohen1999). However, individualism as a cultural pattern emphasizes the independence of the self and individual autonomy (Markus & Kitayama, Reference Markus and Kitayama1991; Triandis, Reference Triandis1995). Collectivism and individualism are very different in the attitude towards family, matrimony (marriage, divorce), individual belief, and social behavior (Triandis, Reference Triandis1994, Reference Triandis1995; Vandello & Cohen, Reference Vandello and Cohen1998, Reference Vandello and Cohen1999).

First, in the collectivist society (region), people enjoy the matrimony and share themselves with their partners in the marriage, and thus they evaluate the self's authenticity at the family level rather than at the individual level (Cross, Gore, & Morris, Reference Cross, Gore and Morris2003; Liu & Chan, Reference Liu and Chan1999). Toth and Kemmelmeier (Reference Toth and Kemmelmeier2009) argue that, in the collectivist society, people always exhibit greater adherence to tradition and social conventions. In China, for quite a long time, collectivism is prevalent and people place the emphases on family unity and self-sacrifice (Singh & Kanjirathinkal, Reference Singh, Kanjirathinkal, Adams and Jones1999). As a result, it is less likely to dissolve the marriage. Comparatively, it is intolerable to sacrifice personal fulfillment for the marriage in the individualistic society. Thus, once the marital relationship let people unhappy, they are highly likely to choose to opt out (Toth & Kemmelmeier, Reference Toth and Kemmelmeier2009). Similarly, Vandello and Cohen (Reference Vandello and Cohen1999) find that states with higher marriage–divorce ratios are more collectivist in the USA. Overall, higher divorce–marriage ratio (divorce rate) is consistently linked with cultural individualism and is more likely in countries with cultural individualism (Dion & Dion, Reference Dion and Dion1996; Lester, Reference Lester1995; Toth & Kemmelmeier, Reference Toth and Kemmelmeier2009). As a matter of fact, the divorce–marriage ratio reflects the contrast between individualistic and collectivistic social cultures, and further higher divorce–marriage ratio, a concrete reflection of the individualistic value or belief, is closely associated with individualism-dominated social culture in a region (society).

Second, individualism-dominated social culture is inversely related with environmentally friendly activities. People in individualism-dominated social culture exhibit lower conformity to group norms and have unfavorable attitude towards public interests (Toth & Kemmelmeier, Reference Toth and Kemmelmeier2009; Vandello & Cohen, Reference Vandello and Cohen1999). Given that environmentally friendly behavior is in the public interests, individualism- dominated regions should exhibit more unfavorable attitude towards environmental protection and have higher likelihood of tolerance for greenwashing. McCarty and Shrum (Reference McCarty and Shrum2001) find that individualism is negatively associated with environmental recycling and sustainable belief. Cho, Thyroff, Rapert, Park, and Lee (Reference Cho, Thyroff, Rapert, Park and Lee2013) find that individualism negatively affects pro-environmental commitment and environmental attitude. More importantly, a branch of very thin but growing literature (Lim & Tsutsui, Reference Lim and Tsutsui2012; Tokar, Reference Tokar1997) provides indirect and preliminary evidence in an incidental way that individualistic (collectivistic) social culture positively (negatively) affects corporate greenwashing. In addition, Roulet and Touboul (Reference Roulet and Touboul2015) find that the positive attitudes towards individualism/liberalism elicit greenwashing.

The sacred marriage requires strong responsibilities and a solemn commitment to stay together for a lifetime (Amato & DeBoer, Reference Amato and DeBoer2001). In this regard, higher divorce–marriage ratio highlights the irresponsible social atmosphere (culture) of breaking promises (Adams & Jones, Reference Adams and Jones1997). As a result, the aforementioned two points, taken together, suggest that individualistic social culture embedded in higher divorce–marriage ratio is positively related with corporate greenwashing.

Lastly, Marquis et al. (Reference Marquis, Glynn and Davis2007)'s community theory states that community isomorphism shapes local social norms and cultural cognition (e.g., the shared local community ideology, value and belief among CEO groups). Bies et al. (Reference Bies, Bartunek, Fort and Zald2007), Gellers (Reference Gellers2012), Pentifallo and VanWynsberghe (Reference Pentifallo and VanWynsberghe2012) conclude that community isomorphism can impel firms and their CEOs to fulfill environmental conservation or involve in greenwashing, depending on whether community isomorphism is individualism dominated or collectivism dominated. Moreover, community isomorphism brings out the ‘peer effect’ (Bursztyn et al., Reference Bursztyn, Ederer, Ferman and Yuchtman2014; Feld & Zölitz, Reference Feld and Zölitz2017), which leads to similar behavior of CEOs within a region (industry). For example, prior literature has validated that the ‘peer effect’ widely exists in CSR and different CSR dimensions (Bollinger, Burkhardt, & Gillingham, Reference Bollinger, Burkhardt and Gillingham2018; Cao et al., Reference Cao, Liang and Zhan2019; Malik, Al Mamun, & Amin, Reference Malik, Al Mamun and Amin2019; Qiu, Yin, & Wang, Reference Qiu, Yin and Wang2016).

The divorce–marriage ratio (divorce rate) is an important social indicator – a numerical measure of portraying a society as a whole (Inglehart, Reference Inglehart1997; Toth & Kemmelmeier, Reference Toth and Kemmelmeier2009). As such, according to the community theory and the ‘peer effect’, higher divorce–marriage ratio is closely associated with individualistic social culture, shapes community isomorphism at the province level, induces the ‘peer effect’ of mitigating environmental involvement, and eventually results in corporate greenwashing. Marriage or divorce belongs to individual-specific event, but high divorce–marriage ratio shapes the individualism-dominated social culture at the macrolevel, which imperceptibly affects all people within a region (province). Maybe CEOs or top managers in some enterprises do not support or agree with individualistic social culture, but they may have to adjust their behavior and corporate decisions due to the community isomorphism, isomorphism pressures, and the ‘peer effect’.

To sum up, higher divorce–marriage ratio characterizes individualism-dominated social culture, which mitigates a firm involving in environmentally friendly activities and increases the extent of greenwashing. Due to the community isomorphism and the ‘peer effect’, the positive effect of individualism-dominated social culture – which is portrayed by divorce–marriage ratio – on corporate greenwashing spreads within a region (province). Thus, we formulate Hypothesis 1 (H1) as below:

The moderating effect of the environmental protection Law (Hypothesis 2)

China has been the largest emerging market and the second largest economy, but rapid economic growth cannot cover the fact that formal institutions fall short and far from adequate (Allen et al., Reference Allen, Qian and Qian2005). As an alternative, informal system (e.g., Confucian culture, social ties) plays the substitutive role in affecting the Chinese society, corporate governance, and individual behavior (Du, Reference Du2016). For this reason, H1 predicts the positive impact of the divorce–marriage ratio – which captures the contest between individualistic and collectivistic social cultures – on corporate greenwashing.

However, it is indisputable that the role of informal institutions depends on how perfect the formal system is (Williamson, Reference Williamson2000). Williamson (Reference Williamson2000) outlines a four-level social institution framework: ‘(1) social embeddedness (e.g., informal institutions, customs, traditions, norms), which remains fairly stable and only evolves slowly over time; (2) institutional environment – formal rules such as judiciary and bureaucracy; (3) governance (e.g., contract); and (4) resource allocation and employment (i.e., price, incentive)’. Moreover, Williamson (Reference Williamson2000) argues that some formal institutions like laws derive from informal system, but formal institution can exert the crowding-out (extrusion) effect on informal system. Extending Williamson (Reference Williamson2000), Elias and Jephcott (Reference Elias and Jephcott1982), Foucault (Reference Foucault2012), Lowes, Nunn, Robinson, and Weigel (Reference Lowes, Nunn, Robinson and Weigel2017) and Tabellini (Reference Tabellini2008) validate that formal institutions may crowd out the original effect from informal systems. Especially, Du et al. (Reference Du, Jian, Zeng and Du2014) find that, the continuous improvement of China's laws weakens the original impact of religious atmosphere around a firm (an informal system) on corporate behavior.

To protect the ecology environment, prevent pollution and other public hazards, ensure public health, promote the construction of ecological civilization and the sustainable development, the Environmental Protection Law of the People's Republic of China has been in force and implemented since 2015.Footnote ² Based on Williamson (Reference Williamson2000)'s four-level social institution analyses framework and the aforementioned discussions, it can be logically inferred that CEPL as a formal institution will crowd out the influence of the divorce–marriage ratio (an informal system and individualistic social culture) on greenwashing. Thus, we formulate Hypothesis 2 (H2) as below:

Sample

The initial sample includes all Chinese listed firms during the period of 2008–2017. Our sample period begins in 2008 because most firms have not disclosed environmental information in 2007 and before. Then, we select the research sample based on several criteria: First, we delete firms pertaining to the banking, insurance, and other financial industries. Second, we discard firm-year observations whose net assets are below zero. Lastly, we eliminate firm-year observations with missing data on firm-specific variables. As a result, we get a final sample of 21,628 firm-year observations, covering 3,003 unique firms. To mitigate the effect of extreme firm-year observations on regression results, all continuous variables are winsorized at the 1 and 99% of their original values.

We sort industries based on the China Securities Regulatory Commission (CSRC) industry classification standard, but manufacturing industry (Coded by C) is divided into ten subindustries, labeled with the second industry code. No serious clustering phenomenon exists except for C4 and C7. Nevertheless, all reported t-(z-) statistics are based on standard errors adjusted for clustering at firm and year level (Petersen, Reference Petersen2009).

Data source

Data sources of variables are reported as below (see Appendix 1): First, the data on corporate greenwashing (GW) are calculated on the basis of hand-collected data on environmental performance by checking annual reports, CSR reports, and corporate official websites (Clarkson, Li, Richardson, & Vasvari, Reference Clarkson, Li, Richardson and Vasvari2008; Du et al., Reference Du, Jian, Zeng and Du2014, Reference Du, Chang, Zeng, Du and Pei2016; GRI, 2006). Second, the data on divorce–marriage ratio (DIV_MAR) are hand-collected from the official website of CNBS (http://www.stats.gov.cn/; China's National Bureau of Statistics). Third, we obtain the data on Marketization index (MKT) from Wang, Fan, and Hu (Reference Wang, Fan and Hu2019). Lastly, we obtain the data on control variables from CSMAR (China Stock Market and Accounting Research).

Corporate greenwashing (dependent variable)

In our study, the dependent variable is corporate greenwashing, labeled as GW. In prior literature, researchers employ two types of proxies for greenwashing (Du, Reference Du2015; Du et al., Reference Du, Jian, Zeng and Chang2018; Laufer, Reference Laufer2003; Ramus & Montiel, Reference Ramus and Montiel2005): (1) ex ante proxy; and (2) ex post proxy. Most extant studies identify greenwashing by ex post proxies such as sustainable ratings (Parguel et al., Reference Parguel, Benoît-Moreau and Larceneux2011) and the lists of greenwashing in the media (Du, Reference Du2015). However, a branch of very thin but growing literature has constructed ex ante proxy to measure greenwashing (e.g., Du et al., Reference Du, Jian, Zeng and Chang2018).

Our measure of corporate greenwashing (GW) can borrow important support from Richardson (Reference Richardson2006), who adopts similar approach to measure overinvestment. Referring to Du et al. (Reference Du, Jian, Zeng and Chang2018) and Richardson (Reference Richardson2006), we construct ex ante proxies for greenwashing: First, we obtain original information from the annual reports, CSR reports, and official websites. Second, we calculate and hand-collect data on a firm's actual value of environmental performance (Appendix 3 and descriptive statistics) based on GRI (2006) and Clarkson et al. (Reference Clarkson, Li, Richardson and Vasvari2008). Third, by regressing Clarkson et al. (Reference Clarkson, Li, Richardson and Vasvari2008)'s model, we obtain the expected value of a firm's environmental performance (Appendix 2). Lastly, we define corporate greenwashing as the difference between the actual value and the expected value of environmental performance if the former is less than the latter, and 0 otherwise.Footnote ³

The divorce–marriage ratio (marital demography, independent variable)

Prior literature adopts three approaches to measure the marital demography (U.S. Bureau of the Census, 1996; Vandello & Cohen, Reference Vandello and Cohen1999): (1) the divorce–marriage ratio (or the reverse scored), measured as the divorce rate at the province level divided by provincial marriage rate ( × 100); (2) the crude divorce rate, computed as the ratio of the number of annual divorces per year to total population; and (3) the fine divorce rate, measured as the ratio of the number of divorces per year to the population of married women, excluding the number of young women below the legal age of marriage.

In this study, we adopt the divorce–marriage ratio (DIV_MAR) based on the first approach for main tests.Footnote ⁴ DIV_MAR is the proxy for marital demography, measured as the divorce rate divided by marriage rate at the province level (U.S. Bureau of the Census, 1996; Vandello & Cohen, Reference Vandello and Cohen1999). If the coefficient on DIV_MAR is positive and significant, H1 is supported.

China's Environmental Protection Law (moderating variable)

In this study, the moderating variable is CEPL. CEPL is an indicator variable, equaling 1 for years of 2015 and after and 0 otherwise. Thus, a significantly negative coefficient on DIV_MAR × CEPL is consistent with H2.

Control variables

To identify the incremental effect of the divorce–marriage ratio on corporate greenwashing, we refer to Clarkson et al. (Reference Clarkson, Li, Richardson and Vasvari2008), Kumar (Reference Kumar2020), and Du et al. (Reference Du, Jian, Zeng and Chang2018) to include a set of control variables into regressions (see Appendix 1 for variable definitions). First, to isolate the impacts of governance mechanisms on greenwashing, blockholder ownership (BLOCK), institutional ownership (INST), managerial ownership (MAN_SHR), the indicator for CEO-chairman duality (DUAL), the ratio of independent directors (INDR), and board size (BOARD) are incorporated into the regression model.

Second, firm-specific financial characteristics such as firm size (SIZE), liability-to-equity ratio (LTE), the ratio of cash flow from operating activity (CFO), corporate growth opportunity (TOBIN'Q), financing activities (FIN), capital intensity (CAPIN), a firm's stock price volatility (VOL) and an indicator for state-owned enterprises (STATE) are included in the regression model. By doing so, the influence of firm-specific financial characteristics on greenwashing can be isolated to some extent.

Third, two macrolevel factors – marketization index (MKT) and GDP per capita at the province level (GDP_PC) – are controlled in the regression model, referring to Du et al. (Reference Du, Jian, Zeng and Chang2018).

Lastly, we include a set of dummies to control for fixed effects of industry and year.

Endogeneity tests using two-stage instrumental variable approach

Next, we adopt two-stage instrumental variable approach to mitigate the potential endogeneity between the divorce–marriage ratio and greenwashing. By doing so, the primary task is to identify the instrumental variable. Spanier (Reference Spanier1984) suggests the negative relation between the number of household population and the likelihood of divorce. Whitehead (Reference Whitehead1998) argues that the divorce is less likely to happen in large-scale families. Referring to Spanier (Reference Spanier1984) and Whitehead (Reference Whitehead1998), the average number of household population (FAMILY) can serve as the appropriate instrumental variable. FAMILY is predicted to be negatively related to the divorce– marriage ratio (DIV_MAR), but it is unlikely that household population affects a firm's greenwashing.

After including FAMILY and all control variables in the first-stage regression, we estimate the fitted value of the divorce–marriage ratio. In column (1) of Table 6, FAMILY is negatively and highly significantly associated with DIV_MAR (−17.5934 with t = −10.21), consistent with our prediction.

Table 6. Endogeneity tests using two-stage instrumental variable regression approach

Note: ***, **, and * represent the 1, 5, and 10% levels of significance, respectively, for a two-tailed test. All reported t-statistics are based on the robust standard errors clustering at firm and year level. All the variables are defined in Appendix 1.

In the second-stage regressions, we further use the fitted value of the divorce–marriage ratio (i.e., DIV_MAR*) as the main explanatory variables. As shown in column (2) of Table 6, the coefficient on DIV_MAR* is positive and significant at the 1% level (.0139 with t = 2.61), authenticating H1 again. In addition, DIV_MAR* × CEPL has a negative and significant coefficient, validating H2.

Overall, after addressing the endogeneity, results in Table 6 suggest that our main findings are not qualitatively changed, and thus our conclusions are less likely to be changed by the endogeneity issue.

Additional tests using individual-level divorce dataFootnote ⁶

Next, we hand-collect individual-level divorce data (DIVORCE).Footnote ⁷ DIVORCE is an indicator, equaling 1 if a firm's CEO, top managers, or the controlling shareholder divorced in the year and 0 otherwise. During our sample period, we only obtain 40 firms with individual-level divorce, so we employ the propensity score matching approach, abide by the one-to-one principle, and adopt ‘ ± .001’ as the threshold to match each DIVORCE (the treated firm) to only one non-DIVORCE (the matched firm). Blockholder ownership (BLOCK), CEO-chairman duality (DUAL), board size (BOARD), firm size (SIZE), financial leverage (LEV), and GDP_PC are chosen to conduct the first-stage matching process. Finally, we get a sample of 80 observations, including 40 treated firms and 40 matched firms.

As Panel A shows, the differences in all chosen variables are significant between the DIVORCE subsample and the non-DIVORCE subsample. However, after conducting the PSM process, all the differences in chosen variables between the treatment sample and the control sample are insignificant. Above results, taken together, suggest that a fairly good work has been done.

As shown in column (2) of Panel B of Table 7, the coefficient on DIVORCE is positive and significant at the 1% level (.9623 with t = 2.87), consistent with H1 and suggesting that firms with the divorced CEO exhibit higher extent of greenwashing than their counterparts. In addition, in column (4) of Panel B, the coefficient on DIVORCE × CEPL is negative and significant (−2.0031 with t = −2.24), verifying H2 again. Overall, results using individual-level divorce data further support H1 and H2.

Table 7. Regression results of greenwashing on individual-specific divorce, environmental law, and other determinants

Note: ***, **, and * represent the 1, 5, and 10% levels of significance, respectively, for a two-tailed test. All reported t-statistics are based on the robust standard errors clustering at firm and year level. All the variables are defined in Appendix 1.