INTRODUCTION
The number of foreign listings has increased fourfold since the early 1980s; indeed, about 10 percent of firms are currently cross-listed in overseas stock exchanges (Fernandes & Giannetti, Reference Fernandes and Giannetti2014; Henderson, Jegadeesh, & Weisbach, Reference Henderson, Jegadeesh and Weisbach2006; You, Parhizgari, & Srivastava, Reference You, Parhizgari and Srivastava2012). However, to ensure that investors are protected from potential managerial wrongdoing, host-country exchanges encourage foreign firms to appoint host-country independent directors to their corporate boards, which signals better monitoring and governance and causes investors to generate significant positive cumulative abnormal returns (Ghosh, He, & Zhao, Reference Ghosh, He and Zhou2015). Surprisingly, researchers have focused on how host-country independent directors prevent firm managers from doing wrong, but not on how those directors prevent managers of other firms from tarnishing reputations, devaluing resources, and diminishing investor confidence, leading to negative spillover effects that transcend national borders, beyond a firm's home country (Fich & Shivdasani, Reference Fich and Shivdasani2007; Kang, Reference Kang2008; Mayer, Reference Mayer2006; McMahon, Reference McMahon2011; Yahya, Reference Yahya2011).
Consequently, we propose two research questions. 1) When foreign firms listed on the host-country exchange are known to submit fraudulent financial reporting in their home country, do other foreign-listed firms from the same home country suffer negative spillover effects on the host-country exchange? 2) Do host-country independent directors reduce spillover effects? If so, how? To develop our arguments, we draw from cognitive categorization and signalling theory research.
To reduce information processing demands, investors tend to cognitively categorize complex objects according to similarities (Kaplan, Reference Kaplan2011; Lickel, Hamilton, Wieczorlowla, Lewis, Sherman, & Uhles, Reference Lickel, Hamilton, Wieczorlowla, Lewis, Sherman and Uhles2000; Mervis & Rosch, Reference Mervis and Rosch1981; Porac & Thomas, Reference Porac and Thomas1994). We contend that when investors are aware that a foreign-listed firm has submitted fraudulent financial reporting, non-errant foreign-listed firms from the same home country will suffer negative spillovers because investors have categorized errant and non-errant foreign firms as similarly likely to report irregularities. In addition, when non-errant and errant foreign-listed firms both have host-country independent directors, investors will perceive that the directors are also responsible for the wrongdoings, leading to greater spillover effects.
Signaling theory indicates that when investors cannot adequately discern the quality of commodities, they will perceive signals of economic value to be more genuine (Certo, Reference Certo2003; Connelly, Certo, Ireland, & Reutzel, Reference Connelly, Certo, Ireland and Reutzel2011; Spence, Reference Spence1973). We contend that when the host-country independent directors of non-errant foreign-listed firms possess home-country, industry, or task-related experiences, investors may perceive that the experience is a credible signal that the directors will fulfil their fiduciary duties, mitigating negative spillovers.
We depart from extant research in two key ways. First, cognitive categorization research has examined negative spillovers among firms located in the same industry or connected by common directors (Porac & Thomas, Reference Porac and Thomas1994; Zavyalova, Pfarrer, Reger, & Shapiro, Reference Zavyalova, Pfarrer, Reger and Shapiro2012; Zuckerman, Reference Zuckerman1999), but not whether investors categorize foreign-listed firms based on their home-country ties or whether director attributes signal information that alters categorizations and mitigates negative spillovers. Furthermore, signalling theory research has examined how leader attributes signal their own firm's legitimacy (Cohen & Dean, Reference Cohen and Dean2005; Higgins & Gulati, Reference Higgins and Gulati2006; Zhang & Wiersema, Reference Zhang and Wiersema2009), but not whether combined director attributes affect spillovers caused by wrongdoings of other firms. Thus, we combine cognitive categorization with signalling theory to better understand factors that influence negative spillovers.
Second, we provide a nuanced view showing that host-country independent directors may fail to reduce negative spillover effects and that inadequate research attention has been paid to contingencies that influence effectiveness (Bell, Filatotchev, & Aguilera, Reference Bell, Filatotchev and Aguilera2014). Because foreign listings are increasingly popular, empirical evidence regarding the effectiveness of host-country independent directors in reducing negative spillovers among foreign-listed firms is essential (Fernandes & Giannetti, Reference Fernandes and Giannetti2014; Henderson, Jegadeesh, & Weisbach, Reference Henderson, Jegadeesh and Weisbach2006; You et al., Reference You, Parhizgari and Srivastava2012). Thus, our results have significant practical implications for host-country exchanges that require foreign firms to appoint host-country independent directors to their corporate boards.
THEORETICAL BACKGROUND AND HYPOTHESES
Cognitive Categorization
Cognitive categorization, in which complex objects are grouped based on distinct similarities, is a core psychological process used to reduce complexity and cope with the strain of information processing (Kaplan, Reference Kaplan2011; Lickel et al., Reference Lickel, Hamilton, Wieczorlowla, Lewis, Sherman and Uhles2000; Mervis & Rosch, Reference Mervis and Rosch1981; Porac & Thomas, Reference Porac and Thomas1994; Rosch, Reference Rosch1975; Shaw, Reference Shaw1990). Strategy research has focused on how upper echelons categorize competitive groups (e.g., Porac, Thomas, & Baden-Fuller, Reference Porac, Thomas and Baden-Fuller2011), perceive strategy, and understand issues (e.g., Dutton & Jackson, Reference Dutton and Jackson1987), but investors also use cognitive categorizations in equity markets when stocks are classified for asset allocation and portfolio management purposes (Barberis & Shleifer, Reference Barberis and Shleifer2003; Peng & Xiong, Reference Peng and Xiong2006). By categorizing stocks based on observable traits, investors can formulate strategies to meet their investment objectives.
Negative Spillovers Through Cognitive Categorization
Although cognitive categorization reduces information processing demands, firms may undertake actions that negatively affect innocent firms (Bourdeau, Cronin, & Voorhees, Reference Bourdeau, Cronin and Voorhees2007; Mayer, Reference Mayer2006). That is, firms can acquire ‘collective reputations’ when investors assume that similar firms share behaviours, based on the premise that individuals are inseparable from their groups, and thus act in ways consistent with current and past behaviours of others in their group (Surroca, Tribo, & Zahra, Reference Surroca, Tribo and Zahra2013; Tirole, Reference Tirole1996). Management researchers have extended this firm-level concept to include industry-level negative spillovers (Barnett & Hoffman, Reference Barnett and Hoffman2008; Barnett & King, Reference Barnett and King2008; King, Lenox, & Barnett, Reference King, Lenox, Barnett, Hoffman and Ventresca2002; Mayer, Reference Mayer2006; Zavyalova et al., Reference Zavyalova, Pfarrer, Reger and Shapiro2012), based on the premise that same-industry firms have similar characteristics and processes (Barnett & King, Reference Barnett and King2008). Thus, when one firm develops a negative reputation, other firms within the same industry may be affected.
Cognitive Categorization and Negative Spillover
National equity exchanges often form common categories for foreign-listed firms with specific country ties. For instance, the New York Stock Exchange categorizes Chinese listed firms as N-shares; the Singapore exchange categorizes them as S-shares. The Singapore market has also created specific indices to track the performance of Chinese listed firms, such as the FTSE ST China Top Index. Group firms from specific home countries are also likely to be categorized as foreign listings with home-country ties (McMahon, Reference McMahon2011; Yahya, Reference Yahya2011). The categorizations influence host-country equity investors (Boyer, Reference Boyer2011; Kumar, Reference Kumar2009) and contribute to reputational penalties and spillover effects within each category. Specifically, when investors receive unexpected information or observe unexpected events affecting a foreign-listed firm, the asymmetric information indicates that non-errant foreign-listed firms will also report irregularities (Kang, Reference Kang2008). Homogenization, in which observers view categories as more homogenous than warranted, may reinforce the negative expectations (Hastie, Reference Hastie1983; Quattrone & Jones, Reference Quattrone and Jones1980).
Homogenization indicates simplified comparisons regarding actors within categories (Bargh, Reference Bargh and Wyer1997; Macrae & Bodenhausen, Reference Macrae and Bodenhausen2000). For instance, stock returns may co-vary beyond fundamental variances according to the index category (Barberis & Shleifer, Reference Barberis and Shleifer2003; Boyer, Reference Boyer2011). In addition, stocks within investment classes tend to show co-moving returns, even when their cash-flows lack common factors (Barberis & Shleifer, Reference Barberis and Shleifer2003). If investors perceive homogeneity among foreign listings with home-country ties, they will perceive that non-errant and errant firms will show similar financial reporting irregularities. They may upwardly revise the probability of irregularities beyond financial reports, assume fundamentally poor governance (Kang, Reference Kang2008), and perceive non-errant foreign-listed firms as more risky, with losses to market value.
Hypothesis 1:
Non-errant foreign-listed firms will, on average, experience negative spillover effects when errant foreign-listed firms announce financial reporting irregularities.
Host-country Independent Directors and Negative Spillovers
Host-country independent directors have fiduciary duties to supervise and monitor management in the interest of firm shareholders. But when they are strongly tied with management, the social norms of reciprocity can cause them to be intentionally complicit in management wrongdoings or to offer only mild and subjective dissent (Stern & Westphal, Reference Stern and Westphal2010; Wade, O'Reilly, & Chandratat, Reference Wade, O'Reilly and Chandratat1990; Westphal & Stern, Reference Westphal and Stern2006; Westphal & Zajac, Reference Westphal and Zajac1997).
Host-country independent directors should add significant value to foreign-listed firms through better monitoring, improved governance, and sharing of superior knowledge about global management practices (Aggarwal, Erel, Stulz, & Williamson, Reference Aggarwal, Erel, Stulz and Williamson2009; Ghosh et al., Reference Ghosh, He and Zhou2015; Giannetti, Liao, & Yu, Reference Giannetti, Liao and Yu2015; Masulis, Wang, & Xie, Reference Masulis, Wang and Xie2012). Being relatively distant from the influence of the home-country management, host-country independent directors are better positioned to fulfil their fiduciary duties (Giannetti et al., Reference Giannetti, Liao and Yu2015). However, they must depend on insiders for information and may lack sufficient information to assess whether insiders are acting according to shareholder interests (Walsh & Seward, Reference Walsh and Seward1990). Thus, significant information asymmetry problems come from the lack of local knowledge about geographically distant home-country institutional environments (Grinblatt & Keloharju, Reference Grinblatt and Keloharju2001; Kang & Kim, Reference Kang and Kim2010; Moore, Bell, & Filatotchev, Reference Moore, Bell and Filatotchev2010).
Information asymmetry problems are likely to be more salient when investors observe irregularities in errant foreign-listed firms. We contend that investors will attribute some responsibility for irregularities to host-country independent directors, leading to cognitive categorizations and homogenizations that increase negative spillovers.
First, investors will perceive that host-country independent directors are partially responsible for wrongdoings committed by errant foreign-listed firms (Kang, Reference Kang2008; Palmrose, Richardson, & Scholz, Reference Palmrose, Richardson and Scholz2004). Investors will focus on the information gaps that prevent directors from fulfilling their fiduciary duties, using the cognitive categorization processes that reduce complexity by assuming collective responsibility.
Second, investors will deduce commonality among host-country independent directors involved in both errant or non-errant firms by homogenizing individuals or groups sharing common characteristics (Shaw, Reference Shaw1990). Thus, investors see mirror effects in the behaviour of host-country independent directors, perceiving that the directors face information gaps that prevent them from fulfilling their fiduciary duties. Indeed, foreign directors often encounter difficulty in staying informed and monitoring management, so that intentional financial misreporting occurs (Masulis et al., Reference Masulis, Wang and Xie2012). Information gaps between various home- and host-country stakeholders were shown to discovery of wrongdoing among Chinese firms listed in foreign markets (Cogman & Orr, Reference Cogman and Orr2013).
In hypothesis 2, we extend the prediction that non-errant foreign-listed firms will suffer negative spillovers because of categorization and homogenization processes. That is, we predict that firms having a higher number of host-country independent directors will experience more negative spillovers.
Hypothesis 2:
The number of host-country independent directors in non-errant foreign-listed firms will increase the negative spillover effects experienced by these firms.
Signalling Value of Host-country Independent Director Attributes
Although categorization and homogenization processes regarding host-country independent directors may increase negative spillover effects in non-errant foreign-listed firms, we suggest that directors may have certain attributes signalling their ability to fulfil their fiduciary duties, reducing negative spillovers.
Signalling theory explains that appropriate signals can counter information asymmetry (Connelly et al., Reference Connelly, Certo, Ireland and Reutzel2011; Spence, Reference Spence1973). Signaling theory has been used to explain how firms can reduce information asymmetries by signaling that they have access to resources, good management, and endorsements. For instance, when managers and directors have previous prestigious employment, investors perceive signals regarding valuable knowledge and resources for ensuring a strong competitive stance (Finkle, Reference Finkle1998; Gulati & Higgins, Reference Gulati and Higgins2003; Higgins & Gulati, Reference Higgins and Gulati2006; Hillman & Dalziel, Reference Hillman and Dalziel2003) and effective organizational actions (Baker & Faulkner, Reference Baker and Faulkner1991; Certo, Reference Certo2003; Cohen & Dean, Reference Cohen and Dean2005; Higgins & Gulati, Reference Higgins and Gulati2006; Suchman, Reference Suchman1995; Zhang & Wiersema, Reference Zhang and Wiersema2009). When young start-ups attract prominent endorsers such as prestigious investment bankers and venture capitalists, they signal quality to potential investors, as prestigious investment bankers would align only with quality firms (Carter & Manaster, Reference Carter and Manaster1990; Higgins & Gulati, Reference Higgins and Gulati2003; Higgins & Gulati, Reference Higgins and Gulati2006; Stuart, Hoang, & Hybels, Reference Stuart, Hoang and Hybels1999).
By adding information symmetry, signals improve performance and ensure congruent categorization and homogenization. When directors have attributes that signal their ability to fulfil fiduciary duties, spillovers may be less likely. That is, when an errant foreign-listed firm announces an irregularity, non-errant foreign-listed firms can signal that their host-country independent directors have sufficient country, industry, and task-related experience to be better, more ethical monitors (Fang, Wade, Delios, & Beamish, Reference Fang, Wade, Delios and Beamish2007; Goerzen & Beamish, Reference Goerzen and Beamish2007; Shaver, Mitchell, & Yeung, Reference Shaver, Mitchell and Yeung1997). Director attributes are disclosed in annual reports of foreign-listed firms and influence how investors perceive the effectiveness of management control, supervision, and ethicality (Filatotchev & Bishop, Reference Filatotchev and Bishop2002; Gulati & Higgins, Reference Gulati and Higgins2003; Higgins & Gulati, Reference Higgins and Gulati2006; Hillman & Dalziel, Reference Hillman and Dalziel2003; Zhang & Wiersema, Reference Zhang and Wiersema2009). Corroborating evidence has shown that firms can dilute negative categorizations if they are also members of positively categorized groups (Vergne, Reference Vergne2012).
Home country experience
We have argued that when errant firms report financial irregularities, investors perceive that non-errant firms are also more likely to do wrong. However, investors may alter their perceptions when the directors of non-errant foreign-listed firms have working experience in the non-errant firm's home country. Foreign-listed firms often come from countries where institutional and cultural environments differ from those in the host country (Kostova & Zaheer, Reference Kostova and Zaheer1999). Consequently, when host-country independent directors know about home-country management rules, they may be better able to recognize when home-country management will commit irregularities (Bierstaker, Reference Bierstaker2009). Rather than rely on top management of foreign-listed firms for country-specific information, host-country independent directors with direct home-country experience can better leverage their knowledge and experience in the home country (Alam, Chen, Ciccotello, & Ryan, Reference Alam, Chen, Ciccotello and Ryan2014; Goerzen & Beamish, Reference Goerzen and Beamish2007; Shaver et al., Reference Shaver, Mitchell and Yeung1997). As a result, investors receive signals regarding monitoring vigilance, transparency, and disclosure (Abada, Encarnación Lucas-Pérezb, Minguez-Verab, & Yagüeb, Reference Abada, Encarnación Lucas-Pérezb, Minguez-Verab and Yagüeb2017; Bierstaker, Reference Bierstaker2009). In addition, being relatively distant and disenfranchised from the home country, the host-country managers can be more critical of potential fraudulent activities, and will avoid wrongdoings that may jeopardize their current and future positions as directors in the host country (Arora & Dharwadkar, Reference Arora and Dharwadkar2011; Fich & Shivdasani, Reference Fich and Shivdasani2007; Homstrom, Reference Homstrom1999). Inferring that the non-errant firm is less risky, investors will alter their categorization and homogenization processes regarding host-country independent directors, mitigating the positive association with negative spillovers.
Hypothesis 3:
Host-country independent directors’ experience in the home country of non-errant foreign-listed firms will moderate the relationship between host-country independent directors and negative spillover effects such that these firms will experience lesser negative spillovers.
Industry experience
Industry experience is similar to home-country experience in altering investors’ perceptions that non-errant foreign-listed firms are likely to do wrong. Directors with industry experience have acquired tacit knowledge about industry conditions and competitive climates and can better monitor managerial decisions and business dealings (Kor & Sundaramurthy, Reference Kor and Sundaramurthy2009). We also contend that managers with industry experience know about industry conditions that pressure, encourage, or enable irregular activities (Zahra, Priem, & Rasheed, Reference Zahra, Priem and Rasheed2005). For example, industries lacking munificence are linked to irregular activities (Baucus & Near, Reference Baucus and Near1991; Staw & Szwajkowski, Reference Staw and Szwajkowski1975) and unethical procurement of resources (Apostolou, Hassell, Webber, & Sumners, Reference Apostolou, Hassell, Webber and Sumners2001). Hence, host-country independent directors with working experience in the industry of non-errant foreign-listed firms may be better able to assess whether industry conditions will create pressures toward irregular activities. Furthermore, given that knowledge of industry conditions is essential to understand firm performance, directors lacking such information may be unable to assess the drivers of performance (Walsh & Seward, Reference Walsh and Seward1990). However, directors who understand industry-specific conditions can evaluate whether financial disclosures are reasonable and consistent with prevailing industry conditions. Thus, investors may rely more on host-country independent directors with working experience in the industry.
Overall, we suggest that host-country independent directors with industry experience will depend less on top management for information, signalling more monitoring vigilance, higher transparency, and disclosure quality (Abada et al., Reference Abada, Encarnación Lucas-Pérezb, Minguez-Verab and Yagüeb2017; Bierstaker, Reference Bierstaker2009). Consequently, investors will perceive that the non-errant foreign-listed firm is less likely to show irregularities. Non-errant firms will no longer be categorized and homogenized with errant firms, mitigating the positive association between host-country independent directors and negative spillovers.
Hypothesis 4:
Host-country independent directors’ experience in the industry of non-errant foreign-listed firms will moderate the relationship between host-country independent directors and negative spillover effects such that these firms will experience lesser negative spillovers.
Task-related experience
Similarly, when host-country independent directors of non-errant foreign-listed firms have task-related experience and financial expertise, investors receive signals that financial reporting irregularities, financial restatements, and accounting manipulations are less likely (Agrawal & Chadha, Reference Agrawal and Chadha2005; DeFond, Hann, & Hu, Reference DeFond, Hann and Hu2005; Firth, Rui, & Wu, Reference Firth, Rui and Wu2011). In addition, corporate boards with financial expertise will be stronger monitors of financial policies and strategies; thus reducing a reliance on incentive contracts for CFOs (Gore, Matsunaga, & Yeung, Reference Gore, Matsunaga and Yeung2011). More important, independent directors with accounting expertise appear to be better monitors of financial reports (Krishnan & Visvanathan, Reference Krishnan and Visvanathan2008).
Thus, firms should appoint independent directors with accounting-related expertise to detect financial reporting irregularities, reduce the reliance on end-of-year financial audits to detect irregularities, and interpret unaudited quarterly financial statements without relying on top management help or influence. Therefore, host-country independent directors with accounting-related experience depend less on home-country managers for information. Consequently, investors see that the independent directors can detect financial irregularities, mitigating the positive association between host-county independent directors and negative spillovers.
Hypothesis 5:
Host-country independent directors with accounting-related experience will moderate the relationship between host-country independent directors and negative spillover effects in non-errant foreign-listed firms such that these firms will experience lesser negative spillovers.
Home-country, industry, and task-related experiences
Home-country, industry, or task-related experiences alone may fail to send sufficiently strong signals (Rediker & Seth, Reference Rediker and Seth1995), but when attributes combine and complement one another, investors may receive more robust signals of monitoring vigilance (Bargh, Reference Bargh1992; Logan, Reference Logan1992). Thus, if each attribute increases perceptions that directors will detect irregularities, categorization and homogenization will be disrupted, further attenuating the positive association with negative spillovers. In support, empirical studies have shown that different types of experience can be complementary in various settings (Colombo & Grilli, Reference Colombo and Grilli2005; Gruber, MacMillan, & Thompson, Reference Gruber, MacMillan and Thompson2012). For instance, industry and business knowledge complements domain-specific accounting knowledge to enhance financial reporting (Dhaliwal, Naiker, & Navissi, Reference Dhaliwal, Naiker and Navissi2010). Investors should respond positively when independent directors collectively possess a complementary combination of home-country, industry, and accounting-related experiences, moderating the relationship between host-country independent directors and negative spillover.
Hypothesis 6:
Host-country independent directors who collectively possess more experience-based attributes will moderate the relationship between host-country independent directors and negative spillover effects such that non-errant foreign-listed firms with these directors will experience lesser negative spillovers than firms with host-country independent directors who collectively possess fewer of these attributes.
METHODS
Sampling
We conducted this research on Chinese firms listed on the Singapore Stock Exchange (SGX). Overseas listings are popular among Chinese firms, but financial irregularities have caused numerous listing suspensions. Over the past decade, China has outpaced other countries in the number of foreign listings in the United States and Singapore (Peng & Blevins, Reference Peng, Blevins, Rasheed and Yoshikawa2012). We studied Chinese firms listed on the SGX for two reasons. First, SGX requires foreign firms seeking listings to appoint at least two independent Singapore residents to the board of directors of the listed entity. Boards must also have a qualified Singaporean to advise the company about local laws or to serve as director resident. Boards must provide ‘negative assurance’ confirmations that financial statements are not misleading or false. These requirements allowed us to avoid sampling foreign-listed firms lacking a host-country or Singaporean independent directors and to ensure that directors have various experience-based attributes. Second, the SGX website provides easy access to corporate governance information from electronic annual reports and official announcements, making Chinese firms listed on the SGX an ideal context for examining negative spillovers.
We avoided using other stock exchanges because corporate governance systems are highly embedded within national systems, so that board independence has varying influence on the effectiveness of corporate governance (Aguilera, Filatotchev, Gospel, & Jackson, Reference Aguilera, Filatotchev, Gospel and Jackson2008; Aguilera & Jackson, Reference Aguilera and Jackson2010; Bell et al., Reference Bell, Filatotchev and Aguilera2014; Bell, Moore, Filatotchev, & Rasheed, Reference Bell, Moore, Filatotchev and Rasheed2012). Thus, shareholders are likely to react differently to Chinese firms listed in Singapore and in other countries regarding announcements of financial reporting irregularities and the impact of governance structures on spillover effects.
To obtain a sample of errant and non-errant foreign-listed SGX firms, we first searched for news articles covering accounting and disclosure irregularities announced by errant foreign-listed firms on the SGX website from 2007 to 2014. The irregularities were severe enough to be covered in major news outlets such as The Straits Times or The Business Times in Singapore. From 2007 to 2012, 17 Chinese firms announced financial reporting irregularities; one in 2007 and 2010, two in 2008 and 2012, five in 2011, and six in 2009. No foreign-listed firms from home countries other than China announced financial reporting irregularities during the sample period.
Second, we used Bloomberg and Osiris to identify 157 Chinese firms that did not announce financial reporting irregularities during the sample period, given that all the errant foreign-listed firms originated from China.
The Measurement of Spillover Effects
To test H1, we used standard financial event study analysis to compute abnormal returns (AR) as a measure of spillover effects on non-errant Chinese firms (McWilliams & Siegel, Reference McWilliams and Siegel1997; Melvin & Valero, Reference Melvin and Valero2009). Operationally, the AR is defined as the actual ex post return on a firm's share price minus the normal return on day t. That is, ARit = Rit − E(Rit), where ARit is the abnormal return on the share price for firm i on event date t, Rit is the actual ex post return on the share price for firm i on event date t, and E(Rit) is the normal return on the share price for firm i on event date t. The normal return, E(Rit), is defined as the expected return if the event of interest (i.e., financial reporting irregularity) had not occurred, and is computed using a market model of the normal share price behaviour. The market model is a statistical model that relates the return of any given share to the return of a specified market portfolio. That is, Rit = αi + βiRmt + εit, where αi is the intercept term, βi is the systematic risk of firm i, Rmt is the rate of return on the Singapore All-Sing Equities Index, which tracks the returns of all SGX-listed equities, on date t, and εit is the error term. To compute the normal return around the event of interest, following McWilliams and Siegel (Reference McWilliams and Siegel1997), we first estimated the market model on a window prior to and not overlapping with the event window in which we expect to observe the AR attributed to the event. We set the estimation window at 500 trading days starting from ten days prior to the event window, and defined the event window up to one day before and after the market reacted to the announcement. Short event windows are preferred because financial event studies assume that markets are efficient and will quickly incorporate financially relevant information into stock prices (McWilliams & Siegel, Reference McWilliams and Siegel1997). The abnormal stock returns for each day in the event window are then summed to arrive at the cumulative abnormal returns (CAR) over the event window. The CARs for each non-errant Chinese firm over the short event windows measure spillover effects. To ensure that the financial event study captures only spillover effects from financial reporting irregularities, we excluded non-errant Chinese firms with confounding announcements one day around the announcement date of financial reporting irregularity (McWilliams & Siegel, Reference McWilliams and Siegel1997). We extracted the stock and index returns for the financial event study analysis from Bloomberg.
Regression Analysis
We used abnormal returns from the financial event study analysis to measure negative spillover effects, the dependent variable.
Independent variables
All independent variables were coded from directors’ profiles disclosed in annual reports supplemented by information from the Osiris database. To test H2, we used the number of Singaporean independent directors, clearly identified in the annual reports and defined by the Singapore Code of Corporate Governance as having no relationship with a firm, its related entities, or its management that could interfere, or be reasonably perceived to interfere, with the exercise of the directors’ objective and independent judgment on corporate affairs. We tested H3 using a dummy variable coded 1 when the firm has Singaporean independent directors with working experience in China as disclosed in the director's profile, 0 otherwise. We tested H4 using a dummy variable coded 1 when the firm has Singaporean independent directors with experience in the industry, 0 otherwise. We used industry codes reported in the Osiris database to determine whether an independent director has working experience in another firm when the director's profile discloses the same industry code as the Chinese firm. We tested H5 using a dummy variable coded 1 when the firm has Singaporean independent directors with accounting-related experience indicated by a professional degree in accounting or work experience in an accounting or auditing firm as disclosed in the director's profile, 0 otherwise. Finally, we tested H6 using a continuous variable that captures the number of experience-based attributes possessed by Singaporean independent directors in each non-errant Chinese firm. The number ranged from 0 indicating no home-country, industry, or accounting-related experience to 3 indicating all three experience-based attributes. We coded all independent variables using directors’ profiles disclosed in annual reports and supplemented by information from the Osiris database.
Control variables
Our empirical model included numerous control variables. First, we controlled for spillover effects through director interlock (Kang, Reference Kang2008) and industry ties (Barnett & King, Reference Barnett and King2008) given that we are concerned with spillover effects from common home-country ties. Director interlock is a dummy variable with a value of 1 if a non-errant Chinese firm has a director in common with the errant Chinese firm, 0 otherwise. Common industry is a dummy variable with a value of 1 if a non-errant Chinese firm is in the same industry as the errant Chinese firm, 0 otherwise. This dummy variable is coded using industry codes from the Osiris database. Controlling for common industry between errant and non-errant Chinese firms is more meaningful when compared with including dummy variables to control for the industries of the errant Chinese firms, considering that industry spillover effects are possible among firms within the same industry (Barnett & Hoffman, Reference Barnett and Hoffman2008; Zavyalova et al., Reference Zavyalova, Pfarrer, Reger and Shapiro2012).
Second, we included variables to control for firm-specific factors. We included annual holding period returns to control for firm performance (Walters, Kroll, & Wright, Reference Walters, Kroll and Wright2010) and market capitalization to control for firm size. We anticipate that negative spillover effects will be weaker for better performing or larger firms because such firms may have better investor relations due to increased visibility in capital markets. We also controlled for systematic firm risk using the capital asset pricing model (Aaker & Jacobson, Reference Aaker and Jacobson1987) as investors may react more negatively for firms with higher systematic risk that cannot be diversified. We extracted stock price data to compute firm performance, size, and risk from Bloomberg.
Third, we controlled for investor familiarity with the listed firm. We expect that greater investor familiarity will lessen negative spillover effects since investors have better access to firm-specific information. We controlled for the number of research analysts using data from Bloomberg and the number of years listed using data from the SGX website. Years listed is computed as the difference between the initial public offering date for a non-errant firm and the irregularity announcement date divided by 365 days. In addition, foreign-listed firms are not required to have a tangible presence in Singapore. Hence, some Chinese firms list in SGX only to raise equity capital; others raise equity capital and are operationally present in Singapore. Furthermore, some Chinese firms may acquire or form alliances with Singapore firms to raise their visibility in the capital market. We included a dummy variable with a value of 1 if a non-errant Chinese firm has an operational presence, acquisition, or alliance in Singapore, 0 otherwise. We coded this dummy variable using the information disclosed in annual reports for operational presence and the SDC Platinum database by Thomson Reuters for acquisitions and alliances. We also included a variable to control whether a firm has a business presence in China. Although all Chinese firms originated from China, they have varying presence in China depending on the extent of diversification into other countries. Chinese firms with a more extensive presence in China will feature more strongly in investors’ cognitive categorization of firms with common home-country ties, thus resulting in stronger negative spillover effects. Annual reports are the best proxy for determining the proportion of active subsidiaries in China over the total number of active subsidiaries, because firms are not required to report geographic segments.
Fourth, we controlled for variables that proxy the strength of a firm's corporate governance to mitigate principal–agent conflicts in accordance with agency theory. In general, we anticipate that negative spillover effects will be greater for firms with poorer governance mechanisms as investors expect poorer monitoring of top executives in non-errant Chinese firms. We computed the dollar value of equity held by Singaporean independent directors and independent directors from China, and entered the natural log of the variable into the model (Bhagat, Bolton, & Romano, Reference Bhagat, Bolton and Romano2008). Independent directors with equity stakes have interests aligned with those of shareholders and thus may be more motivated to monitor top management. The method also controls for discrepancies between countries regarding regulations on the level of equity held by independent directors. We included board size to control for the total number of directors since the monitoring effectiveness of host-country independent directors also depends on how many other directors are on the board. We also controlled for the number of independent directors from China, the home country, as these directors may be better monitors given their familiarity with the home country institutional environment. We included the number of Singaporean independent directors with ties to government entities in Singapore, because investors may view them more positively (Hillman, Reference Hillman2005). We also included the mean board tenure of Singaporean independent directors, as independent directors with longer tenure are better able to monitor top executives, provided social ties with top executives do not compromise their independence over time. We controlled for the number of Singaporean independent directors with board appointments in other non-errant Chinese firms as they may have less time for governing the focal non-errant Chinese firm (Fich & Shivdasani, Reference Fich and Shivdasani2006). Institutional owners may mitigate agency costs (David, Kochhar, & Levitas, Reference David, Kochhar and Levitas1998), so we included the proportion of equity ownership held by government-linked institutional investors owning five percent or more of a firm's equity. We also included a dummy variable, Big 4 audit firm, coded 1 if the external auditor is from KPMG, Ernst & Young, PricewaterhouseCoopers, or Deloitte Touche Tohmatsu, 0 otherwise. Investors may perceive audits by big 4 firms to be more reliable, thus reducing negative spillover effects.
Fifth, we controlled for the impact of family influence in non-errant Chinese firms because family firms may have agency problems in the form of principal–principal conflicts (Filatotchev, Zhang, & Piesse, Reference Filatotchev, Zhang and Piesse2011). We anticipate that negative spillover effects will be greater for non-errant Chinese firms perceived to have higher principal–principal conflicts. Family firms with substantial board and ownership power may appoint and retain host-country independent directors who will not challenge family directors or owners. Indeed, about 35 percent of Chinese firms in the sample have varying degrees of family influence on the board and ownership structure. Hence, we included the number of directors who represent family owners of the non-errant Chinese firm. We also included two other dummy variables to measure family board influence. The first dummy variable has a value of 1 if the board chair is a family representative, 0 otherwise. The second dummy variable has a value of 1 if the founder of the family firm is on the board, 0 otherwise. To control for family influence on ownership structure, we included the proportion of equity held by family members who are also board directors.
Finally, we controlled for year effects via year dummy variables.
RESULTS
Table 1 presents the descriptive statistics for the variables. The financial event study analysis supported H1. Non-errant Chinese firms experienced an average abnormal return of −0.77 percent on the day the market reacted to irregularity announcements (day 0). The negative average abnormal return is statistically significant for the standard parametric and generalized sign tests at p < 0.001. The generalized sign test is more robust to outliers than the standard parametric test (Cowan, Reference Cowan1992), and both test statistics should be significant to support the hypothesis (McWilliams & Siegel, Reference McWilliams and Siegel1997).
Table 1. Descriptive statistics and correlation among study variablesa
Table 1. Descriptive statistics and correlation among study variablesa
Note:
a n = 540 observations. Correlations > |0.08| are significant at p < 0.05 and below.
We tested H2 to H6 using the average abnormal returns at day 0 as the dependent variable, since the generalized sign statistics for average abnormal returns at day −1 (0.30%) and the standard parametric statistic for average abnormal returns at day 1 (0.06%) are insignificant. Standard ordinary least squares regression is not suited for the analyses because repeated observations over the sample period violate the assumption of independence. Furthermore, each non-errant Chinese firm had repeated observations within 2008, 2009, 2011, and 2012 when multiple irregularity announcements occurred. This poses concern about the analysis because the independent variables are invariant within each year, since director information is reported annually. To resolve this problem, we computed the mean abnormal returns from irregularity announcements within each year to derive an average abnormal return from spillover effects for each non-errant Chinese firm. The final dataset comprised 157 unique firms with 540 observations. One to six observations occurred for each non-errant Chinese firm from 2007 to 2012. Firms had six observations if we had complete stock price data to compute the abnormal returns across all six years and no confounding announcements occurred when errant firms announced an irregularity. The dataset is essentially an unbalanced cross-section panel data, which allows us to use a more robust fixed-effects regression model based on the results of the Hausman specification tests. All continuous variables were cantered before creating the interaction terms to test the hypotheses. Heteroskedasticity of the residuals is corrected by replacing the standard errors with the Huber-White robust standard errors (White, Reference White1980). We adopted a one-tailed test for all hypotheses given the predicted direction of the association. Table 2 shows the results of the regression analyses.
Table 2. Regression tests on the abnormal returns of non-errant Chinese firms
Notes: n = 540 observations with 157 unique firms; reported p-values are 1-tailed for the hypothesized main effects and interaction terms, 2-tailed otherwise; * if p < 0.05; *** if p < 0.001.
Model 1 includes only the control variables; Model 2 includes the independent variable; and Model 3 is the full model with all interaction terms and is used to test H2 to H5. We report the R2 (within) numbers for each model as a measure of model fit since we use fixed-effects regression analyses. Multicollinearity poses no problem for Model 3 as the mean variance inflation factor (VIF) is 2.08, and no individual VIFs exceed 10.
For the control variables, although the most direct channels for negative spillovers between firms are thought to be director interlocks (Kang, Reference Kang2008) and common industry ties (Barnett & King, Reference Barnett and King2008), we found no significant effects, perhaps because of the few director interlocks and common industry ties in our sample. As expected, the coefficients of firm performance and firm size are positive but only the latter is significant. It may be that investors do not pay much attention to firm performance since financial reporting irregularities cast doubts on the validity of figures in financial statements. Consistent with studies suggesting that independent directors with equity stakes may be more motivated to monitor top management (Bhagat et al., Reference Bhagat, Bolton and Romano2008), we found that when host-country independent directors have equity ownership, the firms experienced a more positive abnormal return. However, we found no significant effects of home-country independent directors or of their equity stakes on abnormal returns. Interestingly, the coefficient of equity ownership of government-linked institutional investors is positive but only marginally significant. All remaining control variables for principal–agent conflicts are insignificant. Likewise, the controls for family influence to account for principal–principal conflicts produced mixed results. Only the coefficient for family directors is marginally significant, but not in the anticipated direction. It may be that investors have positive perceptions of family influence because of better financial reporting practices in family firms due to reputation concerns and better monitoring of managers. However, all other control variables for family influence are insignificant. These results suggest that agency theory arguments regarding principal – agent or principal – principal conflicts may not always hold for Chinese firms listed in foreign markets, a point we highlight when discussing future research opportunities.
H2 suggests that the number of Singaporean independent directors is positively associated with negative spillover effects affecting non-errant Chinese firms. We used abnormal returns to measure negative spillover effects: more negative abnormal returns indicated greater negative spillover effects and vice versa. Hence, in support of H2, we found a significantly negative coefficient of host-country independent directors (p < 0.001). H3 suggests that Singaporean independent directors’ experience in China moderates the relationship such that non-errant Chinese firms will experience fewer negative spillovers. The hypothesis was supported: a positive and significant coefficient occurred for the interaction between host-country independent directors with home-country experience (p < 0.05). H4 suggests that Singaporean independent directors’ experience in the industry of non-errant Chinese firms will moderate the relationship with spillover effects such that non-errant Chinese firms will experience fewer negative spillovers. This hypothesis is also supported: the coefficient for the interaction between host-country independent director with industry experience was positive and significant (p < 0.05). H5 suggests a moderated relationship with spillover effects in non-errant Chinese firms will occur when Singaporean independent directors have accounting-related experience. This hypothesis is supported, with a positive and significant coefficient for the interaction between host-country independent director with accounting-related experience (p < 0.05). Finally, when we used model 4 to test H6, multicollinearity apparently caused no problem: the mean VIF was 1.84 and no individual VIFs exceeded 10. H6 suggests that when Singaporean independent directors possess home-country, industry, and accounting-related experiences, the relationship between independent directors and negative spillover effects is moderated, such that non-errant Chinese firms with experienced directors experience fewer negative spillovers than firms with independent directors who lack these attributes. Model 4 shows a positive and significant coefficient for the interaction term between host-country independent director and cumulative experience-based attributes (p < 0.001). Hence, H6 is supported.
Additional Analyses
We conducted five additional analyses to check the robustness of the results. First, we tested H1 using alternative event windows. The results continued to support the hypothesis. The CARs for two- and three-day event windows around day 0 remained negative and significant at p < 0.05 or lower: (−1,0), (0,1) and (−1,1).
Second, we selected a normal period before financial reporting irregularities were announced to check for an effect of host-country independent directors on the CARs of non-errant Chinese firms: a (−4, −2) 3-day window. H2 is additionally supported if a nonsignificant coefficient occurs for host-country independent directors during the normal period but a significant coefficient occurs, as reported in Table 2. The coefficient was negative but not significant for the (−4, −2) event window and remained insignificant when we extended the analyses to include (−5, −3) and (−6, −4) 3-day windows. The results additionally support the negative effect of host-country independent directors for non-errant firms.
Third, rather than include all three interaction terms into model 3 at once, we checked whether H3 to H5 were supported if we included each individual interaction term into the model separately. The hypotheses continued to receive support at p < 0.01.
Fourth, we tested H2 to H6 with the dependent variable defined as the cumulative abnormal returns over the (−1,0), (0,1) and (−1,1) event windows. All hypotheses continued to receive support for the (−1,0) window at p < 0.05 and below. For the (0,1) window, all hypotheses received support, except for those on home-country (H3) and accounting-related (H5) experiences (the coefficient signs for H3 and H5 are in the correct direction, but insignificant). For the (−1,1) window, all hypotheses received support at p < 0.05 and below, except for those on home-country (H3) and industry (H4) experiences, which are supported at p < 0.1. The weaker support observed for event windows (0,1) and (−1,1) is likely because the average abnormal return on day 1 is statistically insignificant for the standard parametric test. Since the market has reacted to the irregularity announcements in day 0, including day 1 into the computation of cumulative abnormal returns introduces more noise to the analyses and reduces support for some hypotheses. Overall, the results remain substantively unchanged when we use alternative two- and three-day event windows around day 0 to measure the dependent variable, especially when we exclude the average abnormal return on day 1 from the measurement. Tables 3 and 4 show the results of these robust analyses.
Table 3. Regression tests on the abnormal returns of non-errant Chinese firms with alternative event windows
Notes: n = 540 observations with 157 unique firms; reported p-values are 1-tailed for the hypothesized main effects and interaction terms, 2-tailed otherwise; * if p < 0.05; *** if p < 0.001.
Table 4. Regression tests on the abnormal returns of non-errant Chinese firms with alternative event windows
Notes: n = 540 observations with 157 unique firms; reported p-values are 1-tailed for the hypothesized main effects and interaction terms, 2-tailed otherwise; * if p < 0.05; *** if p < 0.001.
Finally, we note that several control variables in models 3 and 4 of Table 2 have insignificant influences on the dependent variable. We dropped all control variables with p-values above 0.2 for each model and verified that the hypotheses remain supported. We also dropped eight control variables that are theoretically less important (firm risk, analyst coverage, years listed, acquisitions/alliances and operational presence in Singapore, active subsidiaries in home country, host-country independent directors with host-country government ties, host-country independent directors with board appointments in other home-country firms and big 4 audit firms) and confirm that all the hypotheses remain supported.
DISCUSSION
In this study, we examine how host-country investors form cognitive categorizations and homogenizations when foreign-listed firms report financial irregularities. That is, investors assume that non-errant foreign-listed firms are similar to poorly governed, errant foreign-listed firms. Consequently, the negative cognitions will spillover to harm non-errant firms. However, when non-errant firms have host-country independent directors with home-country, industry, and task-related experiences, investors will perceive that the directors will fulfil their fiduciary duties, thus mitigating the positive association with negative spillovers.
We first provide evidence of significant negative spillover effects among foreign-listed firms from China (home country) that are listed in Singapore (host country). The results suggest that due to cognitive categorization, investors are pricing in a higher risk of irregularities occurring in non-errant foreign-listed firms when an errant foreign-listed firm announces a financial reporting irregularity.
Next, we find that when non-errant firms have a great many host-country independent directors, the negative spillover effects increase correspondingly, suggesting that investors are reinforced in perceiving that host-country independent directors of non-errant and errant foreign-listed firms are similar, and that as the number of directors increases, so does the likelihood of wrongdoing. The results suggest that on average, appointing an additional host-country independent director reduces the abnormal return of a non-errant Chinese firm by 7.4%, thus increasing negative spillover effects. The finding corroborates prior research suggesting that host-country independent directors of foreign-listed firms may provide indiscernible benefits (Bris & Cabolis, Reference Bris and Cabolis2008; Dahya, Dimitrov, & McConnell, Reference Dahya, Dimitrov and McConnell2008; Masulis et al., Reference Masulis, Wang and Xie2012).
However, when the directors possess home-country, industry, or accounting-related experience, the association is weakened, probably because their experience signals that they will fulfil their fiduciary duties, which alters prior categorization and homogenization processes in which investors expect errant and non-errant foreign-listed firms to behave the same. Hence, experience-based attributes will mitigate the positive association between host-country independent directors and negative spillovers.
Figure 1 presents graphic representations of the three significant interaction terms in model 3 of Table 2, following procedures suggested by Aiken and West (Reference Aiken and West1991). The x-axis labels of ‘Low’ and ‘High’ represent one standard deviation below and above the mean value of the number of host-country independent directors respectively. The graphs show that the number of Singaporean independent directors is less negatively associated with abnormal returns when the directors have China, industry, or accounting-related experiences. The moderating variables reveal significant effect sizes. When non-errant Chinese firms appoint an additional Singaporean independent director without China experience, the abnormal return is expected to decrease by 3 percent more when compared with appointing an additional Singaporean independent director with China experience. Similarly, the return decreases by 3 percent more for industry experience and 4.4 percent more for accounting-related experience.
Figure 1. Graphic representations of interaction terms in Model 3
Furthermore, we find that when host-country independent directors collectively possess high cumulative experience-based attributes, the association between host-country independent directors and negative spillovers turns from positive to negative. The direction changes probably because investors receive stronger signals that host-country independent directors will fulfil their fiduciary duties, so that investors will abandon their prior cognitive categorizations assuming that non-errant and errant firms have similar poor governance. Figure 2 presents graphic representation of the change following procedures suggested by Aiken and West (Reference Aiken and West1991). The x-axis labels of ‘Low’ and ‘High’ represent one standard deviation below and above the mean value of the number of host-country independent directors respectively. ‘Low cumulative experience-based attributes’ and ‘High cumulative experience-based attributes’ represent one standard deviation below and above the mean value of the host-country independent directors’ cumulative experience-based attributes respectively. This figure reveals an interesting crossover interaction effect in model 4. Specifically, when Singaporean independent directors have low cumulative experience-based attributes, the number of directors is negatively associated with abnormal returns. However, the association becomes positive when they have high cumulative experience-based attributes. Apparently, home-country, industry, and accounting-related experiences are complementary. The finding aligns with research showing that foreign-listed firms from home countries with weak minority investor protection, such as China, benefit by adopting more governance mechanisms because diminishing returns from redundancies and costs of additional governance are no issue (Bell et al., Reference Bell, Filatotchev and Aguilera2014). Hence, by appointing a Singaporean independent director with a complementary experience-based attribute, non-errant Chinese firms will increase abnormal returns by 3.4 percent, about 4.4 times the average 0.77 percent decline in abnormal return they will experience after an irregularity announcement.
Figure 2. Graphic representations of interaction term in Model 4
Contributions
This research makes several contributions to the literature. First, we contribute to emerging cognitive categorization and homogenization research by demonstrating theoretically and empirically that investors categorize and homogenize foreign-listed firms based on home-country ties. Thus, we are the first to link cognitive categorization and homogenization among investors across country boundaries. We also indicate that firms can alter their cognitive categorical membership by focusing on directors’ attributes as resources. Previous research showing how firms can alter cognitive categorization has mainly focused on using labels and language to manipulate categorical membership (Granqvist, Grodal, & Woolley, Reference Granqvist, Grodal and Woolley2013; Kennedy, Reference Kennedy2008; Vergne, Reference Vergne2012). Thus, we offer a new angle, director experience, for manipulating cognitive categorization. Furthermore, we extend research showing that negative spillovers occur mainly through within-industry collective reputations (Barnett & King, Reference Barnett and King2008; King et al., Reference King, Lenox, Barnett, Hoffman and Ventresca2002; Mayer, Reference Mayer2006). Given that irregularities are firm-specific, spillover effects are consistent with our argument that investors categorize and homogenize foreign-listed firms based on home-country ties.
Second, we contribute to signalling theory research by showing that independent directors send both positive and negative signals. Appointing foreign directors to corporate boards may signal commitment ‘to improved corporate governance practices’ (Choi, Park, & Yoo, Reference Choi, Park and Yoo2007; Oxelheim & Randøy, Reference Oxelheim and Randøy2003), but alternatively may cause investors to perceive ineffectively discharged fiduciary duties. Instead, investors must be convinced that directors have sufficient experience-based attributes allowing them to provide substantive oversight. By showing that foreign directors may bring unintended costs, we answer calls for studies showing how directors are more or less valuable (Connelly et al., Reference Connelly, Certo, Ireland and Reutzel2011).
Finally, we extend the empirical discourse on corporate boards and director classifications to better understand how board composition affects monitoring (Johnson, Daily, & Ellstrand, Reference Johnson, Daily and Ellstrand1996). Rather than using the traditional measure of overall director independence, we suggest that information constraints may determine director classifications. Thus, we advance knowledge regarding when host-country independent directors will depend more on top management for information.
Our results have important normative implications for practice. First, we show that investors may engage in cognitive categorization and homogenization based on home-country ties of foreign-listed firms. Thus, foreign firms seeking foreign listings must be aware that negative spillovers may occur within the first few days following revelations of fraudulent activities by firms from their own country, either previously listed or seeking a listing in the same foreign country. In addition, we show that firms may alter investors’ cognitive categorization and homogenization processes by sending signals to assuage concerns that directors will fail their fiduciary duties. Traditional hiring of independent directors may be insufficient. Instead, our data reveal that investors value host-country independent directors that have sufficient experience to monitor top management. Hence, firms are advised to disclose directors’ experience-based attributes in annual reports or other company filings to help individual investors better assess the quality of board monitoring.
Second, our results may interest national stock exchanges and policymakers concerned about the governance of foreign firms. The average abnormal return of −0.77 percent experienced by non-errant Chinese firms listed in SGX is non-trivial. This decline translates into a reduction of about S$1.3 billion in market capitalization for these firms at day 0, or a mean decline of S$2.4 million per firm (median decline per firm is about S$2.2 million). Given our findings of negative spillovers in SGX, we recommend that exchanges, regulators, and enforcement agencies in other countries (including Singapore) evaluate whether mandated governance standards for foreign-listed firms wrongly emphasize director independence with little regard for experience that enable directors to better discharge fiduciary duties (Cohen, Frazzini, & Malloy, Reference Cohen, Frazzini and Malloy2012; Westphal & Zajac, Reference Westphal and Zajac1998).
Limitations and Future Research Directions
Although our study makes several contributions to research on cognitive categorization and signalling theory, our work is like other studies in having limitations to highlight. First, we do not distinguish between different types of investors or reactions to announcements of fraudulent financial reporting, nor do we operationalize cognitive categorization at the investor level. Finance research shows that investors differ in abilities to access and interpret information and thus react differently to positive or negative announcements. Alternative theories should complement and refine the predictions of categorization theory by acknowledging differences in abilities to interpret qualitative and quantitative information (Blau, DeLisle, & Price, Reference Blau, DeLisle and Price2015; Park, Lee, & Song, Reference Park, Lee and Song2014). For instance, naive investors may immediately sell on negative news, whereas more sophisticated mutual funds and insurance companies follow internal company strategies and approval procedures. Thus, future research should avoid treating investors the same, but should compare how different types of investors react after events. For example, the coefficient of equity ownership of government-linked institutional investors, a control variable, may become more significant when a longer event-window is used to account for their reaction time.
Second, we relied heavily on director profiles disclosed in annual reports to code the experience-based attributes of the host-country independent directors. Given the lack of alternative sources of director information, future research could capture other attributes signalling experience-based attributes, such as length of within-country or within-industry tenure.
Third, poor board oversight has been positively associated with financial reporting irregularities, so we focused on board monitoring (Beasley, Reference Beasley1996). Although we demonstrate that experience-based attributes build confidence in host-country independent directors, we still know little about whether investors also form cognitive categorizations according to whether home-country or other foreign-country independent directors have experience-based attributes. Little research attention has been paid to the effects of cultural differences between investors, home-country directors, and other country directors, so a promising topic to consider is how extensively investors rely on home-country or other foreign-country independent directors to monitor the top management of foreign-listed firms. In particular, psychological theories such as social cognition and categorical thinking (Macrae & Bodenhausen, Reference Macrae and Bodenhausen2000) could be applied to study directorships and spillovers. Applying alternative theories to explain investor reactions could be fruitful because several control variables from an agency theory perspective were marginally or not significant. Agency theory may have limited applications given the importance of social relationships for firms operating in China. For instance, social norms of reciprocity may compromise board independence in China. Thus, investor reactions might be sensitive to social influences on governance mechanisms (Ma & Khanna, Reference Ma and Khanna2016).
Finally, we focused on Chinese firms listed on the Singapore Stock Exchange. Although negative spillovers can occur among foreign-listed firms in any country, investors may vary in their cognitive categorization and homogenization processes so that spillover intensities differ across countries. In addition, corporate governance systems are highly embedded within national systems, so signals for altering cognitive categorizations may also vary among shareholders. Future research should capture differences in categorization and signalling strength across countries. Also, we studied whether host-country independent directors have the experience to signal their ability to discharge their duties, but not whether they actually did so. Future research might directly measure their actual decision making and dissent.
CONCLUSION
In conclusion, we have examined how cognitive categorization generates negative spillovers from financial reporting irregularities among foreign-listed Chinese firms in Singapore, and whether signals from host-country independent directors can mitigate negative spillover effects. The findings reveal that host-country independent directors can decrease negative spillover effects when they possess two or more of home-country, industry, or accounting-related experiences. Our study shows how host-country independent directors can use their experiences to protect the reputation of foreign-listed Chinese firms.
