Abstract
This commentary extends the work of Zellweger and Kammerlander by describing how the relationships they hypothesize might be operationalized and tested. It more clearly defines and provides suggestions for how to measure the heterogeneity and complexity of family members and family relationships. It discusses how to operationalize the scope, scale, and complexity of assets controlled by the family, and the characteristics of agents who manage the process. In addition the commentary discusses outcome measures for the family such as family wealth, family communication, family happiness, and family functionality. It describes financial and operational measures of performance for the businesses controlled by the family. The commentary concludes by discussing outcome measures for individual family members and for the agents they employ.
Zellweger and Kammerlander (2015) make the point that most agency–based family business research treats the family as a single homogeneous unit that controls a single firm (Schulze, Lubatkin, Dino, & Buchholtz, 2001; Villalonga & Amit, 2006). However, most families are comprised of individuals with heterogeneous interests, goals, and preferences, and many business families control multiple businesses and other assets. This adds a degree of complexity that generally has not been dealt with by the research community.
The central constructs of the conceptualization presented by Zellweger and Kammerlander (2015) are four control structures that are clearly described and defined: (1) the uncoordinated family, in which control over the management of family affairs is kept in hands of the family without an agent; (2) the embedded family office in which the family appoints a fiduciary from within the existing asset structure such as a treasurer or chief financial officer of a focal family firm to manage the family's financial affairs; (3) the single family office in which families employ an agent with the specific purpose of representing family needs; and (4) the family trust or foundation created to manage family wealth.
Zellweger and Kammerlander (2015) describe a model in which family heterogeneity and asset complexity influence the choice of family control structures which in turn have the potential to influence subsequent outcomes. Their work focuses on the agency conflicts that arise when different control structures are implemented. In this commentary I propose to more clearly identify and define variables that can be used to measure the heterogeneity and complexity of family members and family relationships; the scope, scale, and complexity of assets controlled by the family; the characteristics of agents who manage the process; and a variety of relevant outcomes.
Constructs and Measures for Supporting Variables
As pointed out by Zellweger and Kammerlander (2015), family businesses are controlled by family members with heterogeneous characteristics who in turn manage an asset base that is in most cases more extensive than a single business, and often employ agents or intermediaries.
Heterogeneous Families
The most accepted definition of a family business is one in which family members have both ownership and management responsibilities (e.g., Chua, Chrisman, & Sharma, 1999). This is determined at the “family” level of analysis. Generally, researchers have considered the family as a unitary concept rather than an amalgamation of individual actors who may be involved as owners, employees, managers, and/or board members. Yet families are comprised of heterogeneous members and there are a number of sources of family heterogeneity. With more family members involved there is the potential for greater heterogeneity, thus the number of involved family members that could be determined by a simple count provides a rough indicator of family heterogeneity, but it would not be adequate by itself.
Role Conflict
Kellermanns, Eddleston, Barnett, and Pearson (2008) considered the degree of involvement of individual family members and how it impacted entrepreneurial behavior and performance, demonstrating how consideration of each family member's level of involvement can yield important insights. Family members may be involved in roles as members of the family, as owners of the family assets, and as managers and employees in family businesses. Because of simultaneous roles there can be an intermingling of family considerations and business decisions. These may lead to role conflicts with regards to family issues, while family decisions may be unduly influenced by business needs. Thus, the management of family assets may lack marketplace objectivity and profit discipline, while individual family members may feel that family values are sacrificed for the good of the family assets (Tagiuri & Davis, 1996). Researchers have considered role conflict in family business (e.g., Lee & Rogoff, 1996), and there is a significant body of work regarding work/family role conflict in a broader context (e.g., Carlson, Kacmar, & Williams, 2000). Considering role conflict experienced by individual family members may contribute to a better understanding of the influence of family heterogeneity on asset management and wealth creation.
Affective Conflict
In addition to measuring the role conflict that exists for individual family members, there is also the potential for affective conflict to occur among and between family members. Conflict is described as affective when it becomes emotional and is focused on personal incompatibilities or disputes (Amason, 1996). Researchers have discussed such conflict in family business (e.g., Cosier & Harvey, 1998; Kellermans et al., 2004) and Harvey and Evans (1994) developed a measure of three levels of family conflict. In addition, Amason and others have developed measures of affective conflict that could be modified and utilized to assess family dynamics in the context suggested here.
Investment Preferences
From a slightly different perspective, multigenerational family businesses may have several generations of family members involved. Thus there may be heterogeneity with regards to investment needs and preferences. The literature on investment recognizes that individual investors have differing objectives and expectations regarding risk appetite, asset classes, liquidity, dividends, and time horizon (Maginn, Tuttle, McLeavey, & Pinto, 2007). Those approaching retirement may have a much different risk appetite than those of the succeeding generation that are earlier in their careers. Researchers have discussed generational effects (e.g., Bammens, Voordeckers, & Van Gils, 2008), yet I could find no research that considered these issues among individual family members in this specific context.
Power
Zellweger and Kammerlander (2015) state that the choice of governance structure is influenced by the presence of a powerful patriarch or matriarch. The simplest definition of power is one's ability to exert his or her will in the organization (Finkelstein, 1992) either through rewards or punishment (Keltner, Gruenfeld, & Anderson, 2003). In a meta–analytic review, Keltner et al. show that those who have power, such as the powerful patriarchs and matriarchs described by Zellweger and Kammerlander, are more likely to have positive affect, pay attention to rewards that satisfy their own personal goals, process information rapidly, make snap judgments, and have less inhibited social behavior. In contrast, family members with reduced power are more likely to have negative affect, pay more attention to threats and punishment and others’ interests and goals, demonstrate deliberative reasoning and information processing, and have more inhibited social behavior. These differences also contribute to the heterogeneity of individuals within family businesses.
To measure the differences in power it is necessary to understand sources and degrees of power exercised by the family leader. Several researchers have described and measured sources of power (e.g., Hinkin & Schriescheim, 1989; Finkelstein, 1992; Raven, Schwarzwald, & Koslowsky, 1998). Powerful matriarchs or patriarchs have access to multiple sources of power. Brass (1984) showed that the relative position within workflow, communication, and friendship networks are strongly related to perceptions of influence by employees within organizations. In addition, it could prove useful to measure the centralization of decision making within families and family businesses using measures of centralization such as those proposed be Dewar, Whetten, and Boje (1980). Thus, if a matriarch (patriarch) occupies a key position, and has multiple sources of power such that she controls decision–making, communication, rewards, and punishments, she would be a powerful family leader.
Summary
Families consist of heterogeneous individuals. The number of family members involved, the potential for role and affective conflict, generational effects, and the power exercised by family members could all influence the choice of control structures and subsequent outcomes.
Size and Diversity of Asset Base
Zellweger and Kammerlander (2015) point out that most researchers have focused on a single family and a single business but that asset base size and diversity are likely to influence the choice of control structures and the outcomes. In the United States and the United Kingdom most large firms are widely held. In many other countries, most large firms are part of family business groups, in which the family holds controlling blocks in several publicly traded firms, each of which holds control blocks in yet more publicly traded firms (Morck & Yeung, 2003).
While there is a literature that explores differences between portfolio entrepreneurs and novice entrepreneurs (e.g., Alsos & Kolvereid, 1998; Westhead & Wright, 1998), and some discussion of the creation and maintenance of family dynasties (Jaffe & Lane, 2004), I found no specific operationalizations of how to measure the asset base for families that control multiple business and other resources. As a starting point, I propose that in addition to the monetary value of the asset base, the number of separate business units the family controls should be considered. Additional factors that need to be determined include the other assets controlled by the family and how diverse or related are the different parts of the asset portfolio managed by a family. For example, a family may control many units of a franchise, in which case the businesses are highly related. In contrast, a family may control multiple businesses in diverse industries, both domestically and internationally. Such measures and operationalizations will be necessary to better understand the selection and outcomes of specific control structures.
Agency Costs
As mentioned earlier, Zellweger and Kammerlander (2015) identify four structures for managing family assets: (1) the uncoordinated family, (2) the embedded family office, (3) the single family office, and (4) a family trust or foundation. In all except the first, an agent is involved. The article does an excellent job of pointing out different kinds of agency conflicts and agency costs. Although the types of agency costs are highlighted, the relative power of the agents and principals is also likely to influence outcomes. There is a power differential in an agency relationship any time the principal depends more on the agent than the agent depends on the principal or vice versa (Hill & Jones, 1992). Coff (1999) proposes that access to information, replacement cost to the family, and the cost to the agent if she leaves the relationship all influence bargaining position. Specifically, agents that have idiosyncratic information, agents that would be costly to replace, and agents that have other viable options, are in a position of power and can command greater remuneration, thus increasing the cost for the family. Thus, in addition to a clear understanding of the “types” of agency conflict, the characteristics of the power differential between agents and principals may also influence actions and outcomes (Eisenhardt, 1989).
Relevant Outcomes
The principle outcome considered by Zellweger and Kammerlander (2015) is the creation and maintenance of the financial wealth of the family. Habbershon, Williams, and MacMillan (2003) developed a unified systems perspective of family firm performance. As is common in the field, their research focused on a single family business and the transgenerational development and conservation of wealth was the overall objective. However, to explain the development and conservation of wealth their overall model incorporated the actions and outcomes of the family unit, the business entity, and individual family members. While wealth creation may be viewed as the primary objective for all three entities, there may be other relevant and interesting outcomes.
With regards to the family unit, the creation and maintenance of family wealth is likely to be important to many family members. However, given the heterogeneity of individual family members there are a number of other family outcomes that might be important to consider. For example family communication (Kraut et al., 1998), happiness (Booth & Van Ours, 2008), functionality (Stafford, Duncan, Dane, & Winter, 1999), and the creation and preservation of socioemotional wealth (Gómez–Mejía, Takacs–Haynes, Núñez–Nickel, Jacobson, & Moyano–Fuentes, 2007) may all be relevant outcomes.
With regards to multiple business units, different types of businesses and classes of assets may influence the choice of family wealth management structures. Those structures in turn may have a differential impact on how different assets are managed. Thus, individual business unit performance or individual asset class performance may be relevant outcomes. Researchers have advocated that depending on objectives sought, both financial and operational measures of business performance may be appropriate (Kaplan & Norton, 1996; Wright, Gardner, & Moynahan, 2003). I propose that both types of performance measures might be used when evaluating multiple business units held in a family portfolio.
With regard to individual family members, wealth creation and individual income are obviously relevant variables. However, depending on the perspective of the researchers it may also be useful to evaluate psychological well–being (e.g., Parasuraman, Purohit, Godshalk, & Beutell, 1996) using validated measures such as the Oxford Happiness Inventory, the Depression–Happiness Scale, or the Satisfaction with Life scale (see Hayes & Joseph, 2003).
Finally, because Zellweger and Kammerlander (2015) discuss characteristics of the agent with regard to agency costs, the reward structure and outcomes for the agent may have a significant impact on asset performance and family wealth. From an agency perspective both financial and nonfinancial rewards are linked with agent performance. Recent meta–analyses indicate that both financial and nonfinancial incentives have a positive impact on individual performance (see Stajkovic & Luthans, 1997,2003). The field could be extended by focusing on agent incentives and outcomes.
Conclusion
Zellweger and Kammerlander (2015) present an insightful and thoughtful discussion of agency costs when heterogeneous individuals in families control multiple businesses and assets. It represents a step toward a deeper understanding of an issue whose complexity has been ignored. This commentary is intended to give guidance to future researchers who seek to operationalize the model and the associated constructs introduced in the paper. The commentary proposes several sources of family heterogeneity, describes questions that should be asked regarding asset heterogeneity, discusses power differentials between principals and agents, and suggests that family wealth is only one of a number of relevant outcomes. Together, the paper and the commentary contribute to the literature by clearly explaining the role of embedded family offices, single family offices, and trusts in managing family wealth and provide a strong base from which to launch future research.