Non-family chair and corporate performance

With the prevalence of family firms around the world (La Porta et al. 1999; Claessens et al. 2000; Morck et al. 2000; Faccio and Lang 2002), an increasing number of studies have shed light on how family firms perform. One dimension that is widely acknowledged is family involvement, which, it has been suggested, overpowers family ownership and the use of control-enhancing mechanisms in influencing firm performance (Villalonga and Amit 2006; Andres 2008). With aligned interests between family managers and family owners and the consequent weaker principal-agent conflicts, firms with family CEOs have been shown to have better performance than those headed by non-family CEOs (e.g., McConaughy et al. 1998; Anderson and Reeb 2003; Maury 2006; Sraer and Thesmar 2007; Andres 2008; Cai et al. 2012; Isakov and Weisskopf 2014). However, given the small talent pool among families, family CEOs are inferior in running the business – especially descendant CEOs (Smith and Amoako-Adu 1999; Morck et al. 2000; Burkart et al. 2003; Pérez-González 2006; Villalonga and Amit 2006; Bennedsen et al. 2007; Bertrand et al. 2008; Cucculelli and Micucci 2008; Chang and Shim 2014). Meanwhile, as family CEOs allow family owners to maximize private benefits at the expense of minority shareholders (Young et al. 2008; Jiang and Peng 2011), having a family member as the CEO can be detrimental to firm performance.

Though there are ongoing debates on whether family involvement benefits a firm or destroys firm value, most attention is paid to CEOs while little is known about the effect of the board chair. Board of directors plays a critical role in understanding the impacts of family involvement on firm performance (Zattoni et al. 2015). As the leader of the board, the chair is endowed with substantial power in influencing strategic decisions (Krause et al. 2016) and explains an even larger percentage of performance variance than does the CEO (Krause 2017; Withers and Fitza 2017). Thus, instead of focusing on the CEO, we investigate the role of board chair – the more significant leader. We examine whether and how an unrelated board chair affects firm performance and the conditions under which the costs of appointing a non-family chair may be reduced.

In family firms, reputation capital is highly valued by the controlling families (Berrone et al. 2010; Deephouse and Jaskiewicz 2013; Mullins and Schoar 2016). To build a good family image, family firms are more likely to adopt a strategy of IPO underpricing (Leitterstorf and Rau 2014), provide higher quality accounting information (Martin et al. 2016), and engage in extensive corporate social responsibility (Dyer and Whetten 2006; Berrone et al. 2010). However, the reputation concern is not the same for all family firms (Zellweger et al. 2013; Martin et al. 2016), nor is it the same for firms with family or non-family chairs. The chair is the representative of the firm in charge of communicating with external stakeholders and is often under strong scrutiny from the public (Krause et al. 2016). Having a family member serve as the board chair strengthens the identity overlap between the family and the firm, which induces the family to weigh the costs of reputation loss even though the unfavorable effect is at the firm level. In contrast, when a non-family member is the board chair, the controlling family only has a weak reputation concern as they may partially escape blame for exploitative behaviors (Chen et al. 2015). In non-family-chair firms, the controlling families sit in the shadows. Because of the indirect channel through non-family chairs and the loose ties between the family and the firm, family reputation is only weakly affected – even when the firm suffers from a negative image. The incentive of families to proactively serve in the interests of non-family shareholders is also weak due to the limited impact on the family’s reputation. Thus, with lower costs of reputation loss borne by families for self-interested behaviors, firms with family outsiders serving as board chairs may perform worse than family-chair firms.

We test the hypothesis of the non-family-chair effect by using data from China. Compared with developed markets, investor protection in China is quite weak, with major interest conflicts between the controlling shareholder and minority shareholders (Faccio and Lang 2002). Since the leading concern in family firms is expropriation by the family owners, China provides a good setting to examine the effect of family involvement resulting from the principal-agency problem. In addition, the separate board structure with the positions of chair and CEO held by two different people is relatively common in Chinese firms compared with firms in countries like the U.S. This allows us to document the chair effect and eliminate the simultaneous CEO effect. The significant impact of the Chinese economy and the superior power endowed to board chairs by China’s corporate laws also encourage us to concentrate on China. Furthermore, the core of Chinese commercial practices is “guanxi” for which reputation plays a critical role (Park and Luo 2001). The unique social system and collectivist culture highly values reputation capital and thus makes the costs of reputation loss extremely high for controlling families, which strengthens the role of the board chair.

Using a sample of 1466 non-financial A-share listed firms in China, we find that firms with non-family chairs who have no apparent relationship in blood or marriage with the controlling families perform significantly worse compared to family-chair firms. The negative effects of passive control (i.e., a non-family member being board chair) are reflected in lower accounting performance than firms with active family involvement no matter which performance measure we use. This result is robust to a variety of model specifications and after controlling for potential endogeneity issues using the propensity score matching method, instrumental approaches and a difference-in-difference analysis of chair turnover. We consistently find a negative relation between non-family chair and firm performance.

To verify the argument that controlling families expropriate minority shareholders aggressively when the board chair is appointed outside the family, we further examine the conditions under which the negative effect of a non-family board chair on firm performance is mitigated or worsened. We find that the loss in performance resulting from a non-family appointment is reduced when there is a more developed institutional environment, a higher analyst coverage, more intense scrutiny from the public, stronger monitoring by other large shareholders, or when a firm is controlled by an entrepreneur with multiple children. That is, the negative relationship between the non-family board chair and firm performance is more pronounced when the listed firm is under weak outside monitoring and when the controlling shareholder tends to care less about family business longevity. Those results support the reputation argument – that in cases of weak reputation concern, the incentive of the controlling family to extract private benefits increases and leads to lower performance in non-family-chair firms than in their counterparts. Finally, we show that the documented inferior performance of non-family-chair firms cannot be explained by the talent or agency problem of the CEO, the intensity of family monitoring, the excess control rights held by the families and stock liquidity. The evidence from tests on alternative explanations implies that the negative impact of a non-family board chair on firm performance cannot be dominated by factors other than reputation. Overall, we find that chairs from outside the controlling family are suboptimal to family members in term of firm performance when institutional efficiency is low.

Our paper contributes to the literature that examines the effect of family involvement on family firm performance. The findings on this subject have been mixed. Instead of joining the debate on whether family CEOs purely improve or damage firm performance (e.g., Anderson and Reeb 2003; Maury 2006; Bennedsen et al. 2007; Bertrand et al. 2008; Cai et al. 2012; Isakov and Weisskopf 2014), or how family involvement with strategic control and operational control affects principal-agent conflicts (Luo and Chung 2013), we focus on the unexplored role of the board chair in family firms from the perspective of principal-principal conflict and examine the more important question: Under what conditions will the negative impact of a non-family appointment be reduced or reinforced? The various methods we use to address endogeneity concerns further ensure the robustness of our results. The documented differences in performance between firms with family and non-family board chairs also support the argument on the significant heterogeneity among family firms (Chua et al. 2012) and the importance of board in family firms (Zattoni et al. 2015). By showing that having a non-family chair leads to poor firm performance, our paper also contributes to the increasing number of studies which have investigated the role of a board chair in firms’ decision making (Ma et al. 2013; Chen et al. 2015; Krause et al. 2016; Krause 2017; Withers and Fitza 2017).

Moreover, this paper provides insight into the role of family involvement in emerging countries. Most family firm studies have been conducted in developed countries (Filatotchev et al. 2005), yet firm characteristics such as ownership structure and the institutional environment in which family firms operate are quite different from those in emerging markets – with weak investor protection – which may lead to different conclusions (Miller et al. 2007; Isakov and Weisskopf 2014). Therefore, we extend the work on investigating the effect of family involvement in a specific, important and scarcely explored institutional context: China. We further complement prior papers (e.g., Cai et al. 2012; Amit et al. 2015; Chen et al. 2015; Xu et al. 2015) by using data from China to show that the appointment of a non-family chair can erode firm performance, and the chair effect may depend on the intensity of interest conflicts between controlling families and the minority shareholders.

Finally, this paper also contributes to the literature that seeks to understand the function of informal institutions (e.g., Putnam 1993; Pamuk 2000). Due to the threat of being excluded from a group if they violate social norms, members’ behaviors are partially restricted by moral standards (Elster 1989). Recent studies have indicated that reputation can play a critical role in improving corporate governance with respect to managerial agency problems (e.g., Dyck et al. 2008; Karpoff et al. 2008). By showing that reputation serves as a substitute for weak formal institutions since it inhibits controlling families from expropriating minority shareholders and improves firm performance, our paper complements the previous studies on the reputation effect on a firm.

The rest of the paper is organized as follows. Section “Data” describes the sample selection process and offers summary statistics and the outlines of empirical model. Section “Empirical results” shows the empirical results of the relation between having a non-family chair and firm performance. Section “Endogeneity issues” discusses endogeneity concerns, while Section “Moderating effects of outside monitoring and family structure” explores the conditions under which the non-family chair affects firm performance to different degrees. Section “Alternative explanations” tests alternative explanations and Section “Conclusion” concludes.

In this section, we present the detailed sample selection procedures, summary statistics and the correlation matrix of the main variables in the regression.

In this section, we present the main empirical results from the OLS regression model and perform a few additional tests to further verify our findings.

As illustrated in previous research on CEOs, endogeneity concerns arise when we analyze the effects of family involvement on firm performance (e.g., Bennedsen et al. 2007; Cucculelli and Micucci 2008; Fahlenbrach 2009). The primary concern is reversal causality. The anticipation of bad firm performance might increase the likelihood of the nomination of non-family chairs since the controlling families would like to protect family reputation. Another concern is omitted variables as the time-invariant firm characteristics – such as business culture – might drive our results. In this section, we conduct several tests to address the potential endogeneity between having a non-family chair and firm performance.

In this section, we investigate the conditions that impact the effects of a non-family chair on firm performance. As the appointment of a non-family chair reduces the costs of reputation loss borne by families, the expropriation issues are expected to be severe in non-family-chair firms and to lead to low performance. Therefore, we examine whether the strength of external monitoring moderates the relationship between a non-family chair and firm performance, since outside monitoring increases a family’s reputation concerns and thus mitigates their self-interested incentives (Anderson and Reeb 2003).

The proxies for outside monitoring come from government, analysts and the public. We use the index measuring the development of market intermediaries and the legal environment of Chinese provinces by Fan and Wang (2006) to indicate local institutional efficiency.¹² This index has been used in many studies of China (Jiang et al. 2010; Li et al. 2011; Amit et al. 2015; Jiang et al. 2015). The higher the index, the better the minority shareholders are protected.¹³ Furthermore, as the number of analysts following the firm increases the transparency of firm operations, the incentive of the controlling shareholder to conduct tunneling is weakened. We also use the number of times the controllers are included on the Hurun China Rich List, an annual listing of China’s wealthiest entrepreneurs or businessmen/businesswomen, as the proxy for media, public and government monitoring.¹⁴ The Rich List publishes detailed information about entrepreneurs such as their age, birth place and the major firms they control. After the publication of the Rich List, the media have shown great interest in uncovering negative news about billionaires and their companies (Hong 2004). Moreover, the authorities enforce regulations more strictly for the business people on the Rich List (Firth et al. 2014). Entrepreneurs and the listed firms are closely monitored after they are included on the Rich List.

According to studies on multiple large shareholders (Zwiebel 1995; Laeven and Levine 2008), large shareholders are less likely to be active in monitoring when the distribution of shareholdings is more diversified. We calculate the ownership dispersion by dividing the shares held by the controller over the sum of shares owned by the second to fifth largest shareholders. A lower ratio indicates a more balanced ownership structure. In this case, we expect that non-family-chair firms would be more likely to underperform family-chair firms when there are large shareholders with comparable ownerships than those with a leading controlling shareholder.

Finally, we examine the moderating effect of family size approximated by the enforcement of the one-child policy. The likelihood of family transitions affects the horizon of the family, the importance of family reputation and the intention to pursue family interests at the expense of minority shareholders. Similar to Cao et al. (2015), we use the one-child policy as a proxy for whether the controlling shareholders have a limited number of children and as a proxy for the “pool size” for a family successor. Specifically, we assume that the controllers born after 1960 have only one child, while others born prior to 1960 have multiple children.¹⁵

For the results so far, we attribute the lower performance of non-family-chair firms compared with those with family chairs to the weaker reputation concern and the stronger expropriation by the controlling families. In this section, we explore a few alternative explanations.

First, entrepreneurs are considered as a special group of people who possess valuable skills in firm management and knowledge which are hard to pass on (Morck et al. 1988; Villalonga and Amit 2006; Fahlenbrach 2009). However, through formal meetings in the company or informal communication at the dinner table, members of the controlling family have more opportunities to be in contact with the entrepreneurs and may be more likely to share similar values than non-family members. In this case, the underperformance of non-family-chair firms could be due to the insufficiency of managerial talents possessed by non-family chairs. Meanwhile, as the country of origin of Confucian culture, the degree of trust outside the family in China is low (Xu et al. 2015). Combined with weak investor protection and a far from mature labor market for professional managers (Jiang and Kim 2015), non-family chairs may be suboptimal choices for successors compared with family members, as indicated by the principal-agent perspective (Jensen and Meckling 1976). We first use the chair’s formal education as a proxy for managerial capabilities and further examine the difference in investment efficiency between non-family-chair and family-chair firms. Contrary to our expectations, the univariate test suggests that non-family chairs on average receive significantly higher degrees than their counterparts (3.455 vs. 3.372). The lack of significance in the interaction variable (NFChair×CapEx) in column 1 of Table 11 also suggests that family chairs do not have superior talent in investment compared with non-family chairs. These results indicate that management quality and principal-agent conflicts cannot fully explain the underperformance of non-family-chair firms.

Second, the underperformance of non-family-chair firms might be driven by the lower amount of time and energy allocated to monitoring the agents by the controlling families in those firms. To test this hypothesis, we employ two proxies: the number of firms that the controller controls at the same time (Numfirm) and the controller’s age (Agecon). Under this approach, firms with busier or older controlling shareholders are monitored less by the families and perform worse (see columns 2 to 5). But controlling the Numfirm or Agecon variable in the regression does not reduce the significance of the estimates of the non-family-chair dummy (see columns 2 and 4). The coefficient of NFChair×Numfirm in column 3 is not statistically significant, suggesting that the differences in firm performance exists regardless of the number of firms under control. Contrary to our prediction, the interaction term (NFChair×Agecon) is statistically significant and positive as shown in column 5. The controllers in non-family-chair firms are also more than 1 year younger than those in family-chair firms. The results imply that the view of busy or old controlling shareholders deteriorating non-family-chair firms is not accurate. Thus, our results are not driven by the impact from the chair side or differences in attention that families pay to the listed firm.

Third, previous studies have shown that the divergence of cash flow rights and control rights has a significantly negative effect on firm performance as the controlling shareholders might pursue private interests at the expense of minority shareholders (e.g., Claessens et al. 2002; Faccio and Lang 2002; Fan and Wong 2002; Villalonga and Amit 2006; Amit et al. 2015; Liu et al. 2015).¹⁶ It is possible that the controlling families in non-family-chair firms have higher excess control rights than those in family-chair firms, and this difference may result in performance gaps between the groups. We include the excess control rights dummy (Excon) as an additional explanatory variable in column 6 and its interaction with a non-family-chair dummy in column 7. The results show that the coefficients of NFChair are still negative and significant, and the interaction term is not statistically significant. Hence, chair identity plays an additional role in determining firm performance beyond the divergence of control rights and cash flow rights held by the controlling families.

Finally, earlier studies have found that firms with higher stock liquidity experience less severe expropriation problems while using China’s split-share structure reform (SSSR) as an exogenous shock (Li et al. 2011; Chen et al. 2012).¹⁷ The share reform aligns the interests of the controlling shareholder and minority shareholders to some extent as it converts the non-tradable shares held by the controlling shareholder into freely tradable shares. It is possible that the shares of the non-family-chair firms are less liquid in the stock market compared with those of family-chair firms. To address this issue, we examine the impact of a non-family chair on firm performance using a subsample of firms that experience SSSR and compare the changes in performance for both types of firms. We use a time window of 3 years before and after the reform, and we define the dummy variable, Finish, equal to 1 for years when the share reform is completed (and 0 otherwise). We predict that the interaction variable (NFChair×Finish) would be significantly positive if our results were caused by stock illiquidity. Column 8 shows the results of the difference-in-difference test. The coefficient of NFChair retains its sign and significance, while NFChair×Finish is not statistically significant. Therefore, the difference in stock liquidity is not a major concern.

Despite the critical role of family firms in economic growth, relatively equivalent attention from researchers has not been paid to them – especially those in emerging markets with weak investor protection. In this paper, we investigate the impact of family involvement as a board chair by using a sample of Chinese listed firms from 1999 to 2014. We find that the appointment of a chair who has no apparent family ties with the controlling family deteriorates firm performance. The result is robust to a variety of model specifications and a series of methods to address the potential endogeneity issues. Further analyses show that the negative effect of a non-family chair is more pronounced when the listed firm is under weaker outside monitoring, and the controlling families care less about business longevity. Results from these tests support our hypothesis that the interest conflicts between the controlling families and minority shareholders are more severe in non-family-chair firms due to the lower costs of reputation loss borne by the families. We also confirm that the reputation view is most appropriate and consistent with the empirical evidence.

Our paper could be viewed as providing straight and detailed evidence that non-family chairs hamper firm performance, and – more importantly – the conditions under which the negative impact of non-family-chair appointments could be reduced. Our findings are consistent with Anderson and Reeb (2003), which implies that outside monitoring might be a good method for minimizing family misbehavior in markets such as East Asia. We also find support for arguments by Chen et al. (2015), suggesting that families are more likely to undertake self-interested behaviors such as using related-party loan guarantees in business groups when they are able to escape blame. Our paper contributes to the literature on how family involvement affects firm performance and the literature on the influence of board chairs in firms.

Based on our research, investors should be cautious about board chair appointment decisions in family firms. Sitting in a relatively opaque position behind the “scapegoat” of non-family chairs, the controlling families might extract more private benefits at the sacrifice of other stakeholders. Moreover, the implications of our results might be applied to other settings in which the controlling families have a strong influence on firm decisions while investor protection is weak. Although we document the significant and negative effect of a non-family chair on firm performance, our findings are probably conservative since we only focus on the role of the board chair. With detailed information provided on the family side such as family structure, future research may examine how various types of family involvement (e.g., leadership structure, generation effect) affect corporate decisions in different ways.