3.1. Keyword Network Analysis
We conducted a keyword network analysis based on 1471 keywords to uncover the important keyword lists of the three different network measures for (1) the full sample, (2) the top three journals, and (3) the main theories. The full sample shows the overall trend of corporate governance research related to family involvement. The keyword network analysis for selected journals provides evidence on the differences among important outlets, while the analysis for the main theories can improve our understanding of the important theoretical lenses.
Prior to constructing the keyword network, we refined a list of keywords to address non-standardized keywords. As authors can use different keywords to refer to the same issue or concept, we processed the initial group of keywords by changing keywords with fundamentally the same meaning to the same standard word. The rules for refinement included standardization of multiple words into a single form, removing redundant keywords, removing hyphens, avoiding abbreviations, unifying synonyms, and separating multiple terms in a single keyword. After the standardization, we conducted a keyword network analysis using NetMiner, a network analysis tool that facilitates the analysis of both network and unstructured text data through various analytical data handling strategies.
Table 3 reports the total number of keywords as well as the density, distance, and clustering coefficients of the keyword network. Density measures the proportion of actual connections in a network relative to the number of connections possible. The density of a network relies on the size of the network; a large network is generally less dense. The distance of a network refers to the average number of steps along the shortest paths for all pairs of network keywords (i.e., nodes), revealing the degree of efficiency of information in a network. The clustering coefficient describes the degree of connectivity between neighboring keywords in a network, reflecting the ratio of the number of existing keyword links to the total possible links among keywords. For the keyword network constructed, the density (0.006) and clustering (0.006) coefficients are low, indicating a relatively large and not a highly clustered network. At the same time, long distance (2.595) and the low density of the network structure (0.006) implies that family governance research covers various research ideas or topics.
We followed the network literature to identify lists of important keywords for three different types of network centrality:
degree centrality,
betweenness centrality, and
closeness centrality [
47,
48,
49,
50]. Each network centrality measure captures the different aspect of each keyword in the keyword network. First, degree centrality is the number of word connections that the node has, which reflects the intensity of word usage. A highly ranked keyword has many relationships with other keywords. Second, betweenness centrality measures the extent to which a keyword is located in the paths between other keywords [
51,
52]. Hence, a keyword that scores high on betweenness centrality acts as a point of control that connects keyword groups with different characteristics. Since such a keyword bridges a subset of keywords, deleting it from a keyword network would diminish the coherency of the keyword network. Thus, a keyword can score high on betweenness centrality even though it is not highly connected to other keywords in each cluster of keywords [
28]. Finally, closeness centrality measures the number of keywords, on average, a focal keyword passes to reach all keywords in the network. That is, closeness centrality is the total geodesic distance from a focal keyword to all other keywords [
53]. A keyword that is high on closeness centrality provides a higher level of access efficiency in the network [
54].
Table 4 presents the top keywords regarding the degree, betweenness, and closeness centrality. Given that degree centrality is the number of word connections of a node, the top ten keywords in terms of degree centrality are highly connected to other keywords in the literature and represent the major research interests in the field. The top 10 keywords in terms of betweenness centrality bridge the distinctive research themes. The top 10 keywords in terms of closeness centrality are located in the center of the keyword network. The list of important keywords for different types of network centrality share most of the same keywords, such as “family firm,” “corporate governance,” “family ownership,” “agency theory,” “socioemotional wealth perspective,” “ownership structure,” “ownership,” and “family control.” Unsurprisingly, “family firm,” “family ownership,” or “corporate governance” appeared as the top five keywords because most of the research covers the issue of family governance. Notably, “corporate social responsibility (CSR)” did not appear within the top 10 keywords in terms of degree centrality; it was the last among the top 10 keywords in terms of betweenness centrality and closeness centrality. This result implies that while the use of “corporate social responsibility” is not as intense, it still has power in linking other concepts and we can consider it to be one of the major focuses of research. For instance, Terlaak, et al. [
55] finds that CSR performance depends on the level of family control in business group firms because of socioemotional wealth concerns.
We show five important keywords in the keyword network measures for the top three journals based on the number of published papers in
Table 5. While the analysis above indicates the overall importance of a keyword, three journals also show similarities and differences. In
Family Business Review, “socioemotional wealth perspective” is a very important concept as it appears in all three network measures. As noneconomic factors collectively termed “socioemotional wealth” are often recognized as an important feature that distinguishes family firms from nonfamily firms, it is not surprising that family business research widely adopts the socioemotional wealth perspective in explaining the uniqueness of family involvement and that the non-economic benefits that families pursue in making strategic decisions relate directly to corporate governance issues [
56,
57]. The
Journal of Family Business Strategy also shows similar results. The socioemotional wealth perspective is valued in closeness centrality. In betweenness centrality, “psychological ownership” is an important keyword. Similar to the socioemotional wealth perspective, psychological ownership draws on the psychology literature and explains the unique attributes of family involvement [
58]. Since
Corporate Governance: An International Review has general corporate governance or family firm-related keywords in all three types of centrality, it seems to publish research on a wide variety of topics related to family firms’ corporate governance (such as “crisis” in Minichilli, et al. [
59] and Crespí-Cladera and Martín-Oliver [
60]) to the extent that these research interests cannot form an integrated keyword network. Additionally, agency theory is one of the most important theories in the
Journal of Family Business Strategy and
Corporate Governance: An International Review. Although agency theory as a keyword has many connections with other keywords in the
Journal of Family Business Strategy, it does not necessarily bridge separate research themes, nor is it located in the center of the keyword networks.
Table 6 presents the top keywords associated with each of the main theory clusters. Agency theory, stewardship theory, behavioral agency theory, and institutional theory are the theoretical frameworks that are most frequently adopted in family firm research [
61,
62,
63,
64,
65]. Within each theoretical framework, the 10 search keywords most commonly associated with each theory were identified [
35]. Few notable points are worthwhile to mention. First, our analysis shows that research that utilizes agency theory tends to be associated with stewardship theory. Although both agency theory and stewardship theory are often applied to explain the unique aspects of family firms, the two theoretical frameworks are considered as distinctive and opposing approaches to understanding the governance relationships and behaviors in family businesses [
61]. Second, keywords related to innovation and entrepreneurship tend to be associated with studies that adopt stewardship theory as its main theory. Even though family firms are often perceived to be conservative, they represent a surprisingly large portion of the most innovative firms around the world [
66]. As strategic decisions regarding innovation and entrepreneurship require a long-term perspective, studies explain this behavior by the presence of family members that act as stewards of the firm [
67]. Third, keywords related to the management of family firms (e.g., CEO succession, nonfamily CEO, professionalization, CEO compensation) are associated with research adopting behavioral agency theory. Given that the behavioral agency theory predicts that family members’ concerns for socioemotional wealth may lead to risk aversion especially with respect to strategic choices threatening family control [
57,
68], governance issues about CEO succession are a critical strategic decision because family members are known to possess a strong desire to maintain family control, which leads to intention for transgenerational control [
69]. Fourth, keywords related to corporate governance issues that are particularly relevant to emerging economies tend to be associated with studies adopting institutional theory as the main theoretical framework. Since family firms are prevalent in emerging markets which is characterized by weak institutional environments [
70], the presence of family members as majority shareholders creates unique governance problems such as principal-principal conflicts.
3.2. Citation Network Analysis
We conducted citation network analysis to identify both the popularity and relative importance (i.e., prestige) of an article according to the number of citations of the paper received by other highly cited papers [
36,
71,
72]. Based on the assumption that the connections between papers in the citation network constitute channels through which scientific knowledge is transformed [
73], citation network analysis allowed us to identify key papers that are particularly influential in the family governance literature [
74].
Among the 947 papers, 908 papers formed a meaningful citation network (Refer to
Table 7). This result implies that papers in our sample are a good representation of the research on family governance. For the citation network constructed, the density coefficient (0.010) is low, the distance coefficient (3.517) is high, and the clustering coefficient (0.248) is high. This indicates that the citation network is relatively sparse, relatively large, and highly clustered.
In order to overcome the limitations of the citation analysis, we analyzed the citation network using the PageRank algorithm. The PageRank score of a paper in a network comprised of
N papers was computed with the following equation, where paper
A is cited by
n papers (i.e.,
T1, …,
Tn) and parameter
d is a damping factor that refers to the fraction of random walks that continue to spread along with citations. Following the Google PageRank algorithm by Brin and Page [
44], we set the parameter
d to 0.85.
Table 8 lists the top five influential papers according to their PageRank score in the citation network of 908 papers. These influential papers identified by the Page Rank scores are important not only because of their frequent appearance on the reference list, but also because of their higher centrality in the citation network as they connect other important papers as well [
71]. Morck, et al. [
75] examines the relationship between family ownership and market valuation of the family firm, while Anderson and Reeb [
7] examine the relationship between founding-family ownership and firm performance. Other influential papers provide significant evidence of family ownership in not only developed nations such as the U.S. but also in various national contexts such as East Asia [
3,
76,
77].
Lastly, we conducted a clustering analysis for the most cited publications. In order to include publications that are considered to be most cited among our sample, we select papers with an in-degree centrality value of over 0.2. This clustering analysis was conducted using a list of publications that authors need to consider in order to increase the validity of the research idea.
Table 9 presents the main clusters that are identified.
Cluster 1 mainly contains descriptive articles that explore the phenomenon of family governance and its implications for firm value. La Porta, Lopez-de-Silanes and Shleifer [
3] compares the ownership structure of the largest publicly traded firms from 27 countries. Claessens, Djankov and Lang [
77] examine the ownership and control patterns in 9 East Asian countries, while Faccio and Lang [
5] examine that of corporations from 13 Western European countries. These three papers in cluster 1 focus on cross-national differences in the ownership and control patterns and identify family ownership as both prevalent and substantial. The remaining papers explore when family governance creates value for the family firm. Anderson and Reeb [
7] find that not only do family firms perform better than nonfamily firms, their performance is likely to better when the firm is also managed by a family member. Villalonga and Amit [
78] find that family ownership creates value when the founder is the CEO or the chairman of the board; interestingly, firm value is lower when a family descendant serves as CEO. Therefore, cluster 1 validates the phenomenon of family governance to be both prevalent and substantial around the world and its value-creating implications.
Cluster 2 focuses on the underlying factors beyond explicit ownership or control that contribute to family firm sustainability. Astrachan and Kolenko [
79] suggest that the adoption of human resource management practices and professional governance practices is critical for the sustainability of the family business. Litz [
80] proposes that the extent to which family members strive to achieve “intra-organizational family-based relatedness” is important for family firms’ sustainable development. Finding that family CEOs are willing to accept lower executive compensation, Gomez-Mejia, et al. [
81] emphasizes that altruistic family motives–which also become stronger at higher levels of family ownership concentration–contribute to family firm sustainability. Drawing on agency theory, Chua, et al. [
82] find that relationships with nonfamily managers are important for sustaining family businesses. In a study of family firms that failed to achieve sustainability, Miller, et al. [
83] outlines factors such as intergenerational family dynamics and inappropriate leadership styles as key sources. Overall, cluster 2 is interested in identifying various factors other than the concentration of ownership and/or management within the controlling family that contribute to the sustainability of family businesses.
Cluster 3 either compares different types of family businesses or family firms with their nonfamily counterparts in order to delineate the uniqueness of family governance. Focusing on business strategies, Daily and Dollinger [
84] suggest that family firms that are managed by family firms tend to adopt a more conservative approach by avoiding growth-oriented strategies compared to their counterparts that are professionally managed. Schulze, et al. [
85] draw attention to the unique agency relationships in family businesses that are characterized by altruism. Even if family agents are given performance-based incentive pay, family businesses have been found to experience agency problems that are difficult to mitigate [
85]. Miller, et al. [
86] investigates a more fine-grained measurement of family business (i.e., lone founder businesses vs. true family businesses) to provide a better understanding of whether family governance is beneficial or harmful for its sustainability; specifically, authors emphasize that market valuations depend greatly on how family businesses are defined. Gómez-Mejía, Haynes, Núñez-Nickel, Jacobson and Moyano-Fuentes [
57] identify the loss of socioemotional wealth as the primary reference point that family businesses use to make decisions.
3.3. Author Network Analysis
Based on a list of 1670 authors that contributed to research on family governance in our sample of papers, we aimed to determine the important players are in the overall research stream by analyzing authors’ names in the citations. We analyzed the authors’ publication frequency and connections through the references. As
Table 10 shows, the author network is relatively large but highly connected. The long distance (5.373) along with the high clustering coefficient (0.871) implies that while there is a highly connected network of authors that cover family involvement issues frequently, the sample contains many authors who published papers on family involvement less frequently.
In order to better understand the author network of the family governance literature, three network-related measures (i.e., degree centrality, betweenness centrality, and closeness centrality) were examined.
Table 11 provides the top five contributors to the research stream in terms of frequency and each of the network-related measures. The top five influential authors by the number of publications are Miller D., Kellermanns F.W., Chrisman J.J., Le Breton–Miller I., and De Massis A. These top five authors contribute to nearly nine percent of the literature. The top two papers by the number of citations published by Miller D. with Le Breton–Miller I. are “Are family firms really superior performers?” published in the
Journal of Corporate Finance, and “Why do some family businesses out-compete? Governance, long-term orientations, and sustainable capability” published in the
Entrepreneurship:
Theory and Practice. The focus of Miller D.’s family governance research provides meaningful evidence on the innovative outcomes and financial performance of family firms and the strategic choices of family founder/CEOs.
The analysis of the top five authors based on the three network measures shows the important presence of authors in the research stream. For all three centralities, Chrisman J.J.’s papers have the highest centrality scores. The two important works by the number of citations are “Succession and nonsuccession concerns of family firms and agency relationship with nonfamily managers” published in the Family Business Review and “Are family managers agents or stewards? An exploratory study in privately held family firms” published in the Journal of Business Research. Chrisman J.J. publishes important papers that mainly deals with the relationship between family ownership and financing, family succession, and analytic views on family business literature. Chirico F., Nordqvist M., and Chua J.H. are not in the top five by frequency, but they have an important presence in terms of their connections with the research community. Among Chirico, F.’s work, “Family firm heterogeneity and governance: A configuration approach” published in the Journal of Small Business Management provides nine categories of family firm types that are useful in reflecting on the meaningful differences in ownership and management of family firms.