What

Agency theory is one of the most widely used theories in management (Daily et al. 2003; Wasserman 2006), and family firms in particular (Madison et al., 2016). Broadly, agency theory is about the relationship between two parties, the principal (owner) and the agent (manager; Eisenhardt 1989; Jensen and Meckling 1976; Ross 1973). More specifically, it examines this relationship from a behavioral and a structural perspective. Theory suggests that given the chance, agents will behave in a self-interested manner, behavior that may conflict with the principal’s interest (Chrisman et al. 2004; Eisenhardt 1989; Jensen and Meckling 1976; Wiseman et al. 2012). As such, principals will enact structural mechanisms that monitor the agent in order to curb the opportunistic behavior and better align the parties’ interests (Cruz et al. 2010; Eisenhardt 1989; Fama and Jensen 1983).

How

Firm performance by way of cost minimization and greater efficiencies is the desired outcome of the agency theory perspective (Corbetta and Salvato 2004; Fama 1980). Theory suggests that agency costs are incurred to alleviate the agency problems associated with the separation of the firm’s ownership and management (Eisenhardt 1989; Jensen and Meckling 1976; Karra et al. 2006; Lee and O’Neill 2003; Wasserman 2006). Separation of ownership and management is a key component of agency theory; the principal delegates work to the agent, and the agent is expected to act in the best interest of the principal (Ross 1973; Wiseman et al. 2012). However, when the interest of the principal is not aligned with the interest of the agent, and when the principal lacks the information to assess the agent’s behavior, an agency problem is created (Eisenhardt 1989; Karra et al. 2006; Lee and O’Neill 2003; Ross 1973). Agency problems are categorized as moral hazard or adverse selection (Chrisman et al. 2004; Eisenhardt 1989; Karra et al. 2006). A moral hazard agency problem is when the agent lacks effort in the scope of the employment relationship (Chrisman et al. 2004; Ross 1973). It is considered a form of opportunistic behavior that includes free-riding, shirking, and perk-consumption (Chrisman et al. 2004; Chua et al. 2009; Karra et al. 2006). An adverse selection agency problem is when the agent lacks the ability or skills to competently behave in the scope of the employment relationship (Eisenhardt 1989; Fama 1980; Schulze et al. 2001).

Agency problems can be reduced in two ways (Eisenhardt 1989). The first is to create a governance structure that enables the principal to monitor and assess the actual behavior of the agent (Anderson and Reeb 2004; Chrisman et al. 2007). This structure includes, for example, reporting procedures, additional management, or a board of directors (Donaldson and Davis 1991). The second is to create a governance structure where the employment contract is based on actual outcomes (Eisenhardt 1989). For example, compensation incentive plans, where the agent receives incentives for high firm performance, fall into this category (Chrisman et al. 2007). This shifts the risk to the agent, thereby creating alignment between the principal and agent’s interests (Davis et al. 1997; Eisenhardt 1989). In sum, the principal can choose to establish a governance structure based on the agent’s actual behavior or the outcomes of that behavior (Eisenhardt 1989). Either choice creates agency costs for the principal (Jensen and Meckling 1976).

Why

The underlying assumption of agency theory is based on the economic model of man (Davis et al. 1997; Eisenhardt 1989; Jensen and Meckling 1976). This model assumes that individuals seek to optimize their own utility. In the principal-agent relationship, an agent is hired to maximize the principal’s utility (Ross 1973). However, agency theory assumes agents will instead behave opportunistically because they too are self-serving. Therefore, the principal enacts mechanisms to minimize losses to their own utility (Davis et al. 1997; Eisenhardt 1989; Jensen and Meckling 1976; Ross 1973).

Where

As described, agency theory contains the essential elements of theory. It has predictive and explanatory power as shown in the descriptions of the what, how, and why. However, the question remains as to whether agency theory is adequate and relevant when context is taken into consideration. In other words, are expansions necessary or boundary conditions or limitations apparent if the where is considered? To assess its strength and generalizability, the next section of this chapter examines agency theory within family firms, an important organizational context. Family firms represent a unique context for several reasons. Family-controlled firms differ from their nonfamily controlled counterparts along several important dimensions. On the one hand, families have little need to monitor management as their equity stakes overlap considerably with their own. On the other, large equity stakes give families considerable discretion to implement strategies that serve the family, providing them with incentives to monitor management more closely and the voting power to discipline them should they engage in unproductive, opportunistic activities.

Characteristics of the Literature

This search process yielded 89 articles grounded in agency theory within the context of family firms. These articles appeared in 22 journals across several disciplines and were published between 2000 and 2014. The articles comprised 18 conceptual articles and 71 empirical articles. Four of the empirical articles were qualitative in nature, and 67 were quantitative. We focus specifically on the quantitative articles for this review because we are interested in the methodological approaches associated with these articles. Table 3.1 provides an alphabetical list and the characteristics of these 67 articles. The focus of our literature review is twofold; we first synthesize the content of agency theory articles, and second we analyze the methodology. We then present future research suggestions that can further extend theory while simultaneously addressing the methodical limitations in extant investigations.

Content of the Literature

Agency Problems. At the heart of agency theory is the separation of ownership and management and the problems that arise from this separation (Eisenhardt 1989). This owner-manager conflict is referred to as a Type I agency problem (Villalonga and Amit 2006). However, it was assumed that family firms are not susceptible to Type I agency problems due to a lack of separation among owners and managers (Chrisman et al. 2004; Goel et al. 2012; Jensen and Meckling 1976); and therefore, agency perspectives remained absent in this organizational context. However, the work of Schulze and his colleagues challenged this logic (Schulze et al. 2001; Schulze et al. 2003a, 2003b), and began an important stream of literature. This stream seeks to conceptually and empirically demonstrate that family firms are indeed susceptible to agency problems. Unique to the family firm, research supports nontraditional agency problems such as those created by asymmetric altruism (Moores 2009; Schulze et al. 2001; Schulze et al. 2003a, 2003b) and entrenched family ownership (Block 2012; Moores 2009; Nicholson 2008).

Altruism is often regarded as a selfless other-serving behavior but is presented in a different light by agency theorists (Eddleston et al. 2008). Agency theory perspectives refer to altruism as asymmetric, which describes behavior that is exploitable and not reciprocated (Chua et al. 2009; Wright and Kellermanns 2011). It can cause harm to the family firm by creating both moral hazard and adverse selection agency problems (Schulze et al. 2001; Schulze et al. 2003a). Moral hazard problems are created by asymmetric altruism when parents are overly generous to their children and their children take advantage of this act by shirking or free-riding in the employment relationship (Dawson 2011; Eddleston et al. 2008; Schulze et al. 2001, 2003a). This agency problem then becomes worse because family firm leaders don’t often monitor the behavior of other family members in the firm (Chua et al. 2011; Eddleston et al. 2010). Adverse selection agency problems are created when family firms hire family members who may be less qualified and skilled than nonfamily members for the position (Karra et al. 2006; Schulze et al. 2001, 2003a), and may even pay them more generously (Chua et al. 2009). These agency problems created by asymmetric altruism are shown to decrease the performance of the family firm (Eddleston et al. 2010; Wright and Kellermanns 2011) and to create the need family firms are burdened with to incur agency costs by implementing governance mechanisms to monitor and assess behavior (Chua et al. 2009; Lubatkin et al. 2007).

Entrenched family ownership can also create unique family firm agency problems. Entrenchment is defined as “the relational contract between owners and managers that enable both to occupy key positions in the firm for a significant duration” (Moores 2009: 172). Family ownership is often described as effective because it reduces Type I agency problems associated with the separation of owners and managers (Anderson and Reeb 2003b; Chirico et al. 2011; Jensen and Meckling 1976; Tsai et al. 2006). However, some research suggests that family ownership is not effective because family dynamics and conflicts are difficult to monitor. With ineffective monitoring, moral hazard agency problems may increase, inhibiting productivity (Block 2012). Furthermore, the emotional attachment the owning family has on the firm may inhibit sound business decisions, decreasing firm performance further (Nicholson 2008).

Additionally, family ownership can create agency problems specific to family firms, referred to as Type II agency problems (Ali et al. 2007; Goel et al. 2012; Villalonga and Amit 2006). Instead of the Type I owner-manager conflict, the Type II represents the conflict between majority and minority shareholders (i.e., family and nonfamily shareholders; Villalonga and Amit 2006). An example of a Type II agency problem manifests itself in the misalignment of shareholder goals; family firms often pursue noneconomic goals at the expense of financial gain (Gómez-Mejía et al. 2007). Diverting resources to pursue the family’s noneconomic agendas may negatively impact firm performance (Chrisman et al. 2004), thus creating conflicts between family and nonfamily shareholders.

In conclusion, the literature provides support that agency problems are indeed prevalent in the family firm context and is important for two reasons. First, this literature challenged the original agency theory research that argued agency problems did not exist in organizations typified by the convergence of ownership and management (i.e., Eisenhardt 1989; Fama and Jensen 1983; Jensen and Meckling 1976). In doing so, a host of nontraditional agency problems specific to family firms (i.e., asymmetric altruism, family entrenchment, Type II agency problems) was brought to light, thus expanding agency theory into the realm of the family firm. Second, this realization allows for continued agency research within a family firm context, but with a different focus. It allows scholars to research whether other tenets of agency theory are applicable to family firms. Accordingly, the next section shifts the focus away from agency problems and instead focuses on the governance mechanisms agency theorists call for to mitigate agency problems.

Agency Governance. Research supports agency theory’s applicability within family firms by examining the firm-level governance mechanisms and their relationship with firm performance. For example, research investigates boards of directors (Anderson and Reeb 2004; Braun and Sharma 2007; Gersick and Feliu 2014; Goel et al. 2014), incentive compensation (Chrisman et al. 2007), and monitoring activities (Chrisman et al. 2007) as agency mechanisms used to mitigate agency problems and subsequently increase performance. Anderson and Reeb (2004) examine board independence and family influence to support their contention that an agency lens is applicable to family firms. They conclude that monitoring mechanisms such as a board of directors are necessary; outside board members monitor the family, and family board members monitor the business (Anderson and Reeb 2004). Likewise, Chrisman and his colleagues (2007) explore the motivations and control of family firm managers to determine if they are agents or stewards. They conclude that family managers are agents because family firms using governance mechanisms prescribed by agency theory (i.e., monitoring, incentive compensation) have higher levels of performance (Chrisman et al. 2007). Braun and Sharma (2007) explore CEO duality and demonstrate that the separation of positions is a beneficial governance structure for family firm performance.

Instead of a focus on performance, others have focused on perceptions. Cruz and her colleagues (2010) examine how agency contracts are affected by the CEO’s perceptions of the top management team. In line with agency theory, they suggest that monitoring and incentive mechanisms are implemented based on the CEO’s trust perceptions of top managers. In the family firm context, they suggest that the presence of top managers who are related to the CEO and the level of family ownership in the firm affect these perceptions. Dawson (2011) examines the perceptions of private equity firms as they assess the attractiveness of family firms. She finds that investors prefer professionalized family firms, described as businesses that have agency prescriptions in place such as formal human resource management practices and the presence of nonfamily managers (Dawson, 2011).

In conclusion, this group of family firm articles focuses on the remedies prescribed by agency theory that are theorized to positively impact firm performance. Importantly, this research provides support that agency governance mechanisms, such as the presence of a board of directors, incentive compensation plans, and monitoring activities serve their theorized purpose within family firms. Similar to agency problem research, this research expands the boundary conditions of agency theory by supporting its governance mechanisms and related outcomes within a new organizational context.

Methods of the Literature

As suggested, the cornerstone of much of the agency theory research is that family involvement in the firm creates or causes unique agency conditions that influence the subsequent performance of the firm. Given the importance of research methods in establishing and validating these causal inferences, we examine the extent to which agency theory research includes research designs that effectively guard against key threats to causality and confounded results. Causal relationships can be inferred when three conditions are met. First, the cause and effect variables covary (Van de Ven 2007). Second, the cause precedes the effect in time (Van de Ven 2007). Third, alternative explanations for the covariance between cause and effect variables are eliminated (Cook and Campbell 1979). While exhaustively meeting these criteria is impossible, researchers can more confidently infer causal relationships as these conditions are addressed more thoroughly with a line of solid evidence being preferred over several lines of evidence of questionable quality.

The first criterion, covariation between cause and effect variables, suggests that a change in one variable is accompanied by a change in another (Van de Ven 2007). Presenting a correlation matrix that includes all the study variables provides a rather straightforward means to demonstrate covariance. Given the importance of covariation to causality (Van de Ven 2007), these matrices would be expected to be ubiquitous; however, reviews of the entrepreneurship literature have indicated that not all studies report these matrices (e.g., Crook et al. 2010). Indeed, we find this to be the case in our review of the family firm agency theory literature, finding that only 48 studies (71.6 percent) reported a covariation matrix.

Second, it is critical to establish that the cause precedes the effect, suggesting that the change in the cause variable occurs before the change in the effect variable (i.e., the agency conditions cause changes in outcomes; Van de Ven 2007). The ability to test whether an effect is actually caused by the hypothesized variable is limited when researchers rely on cross-sectional data where the cause and effect variables are collected at the same point in time. Hence, from a temporal perspective, causal influences can be inferred more confidently with longitudinal data or, at the very least, cross-sectional data using time-lagged variables. In our review, most of the investigations are cross-sectional in nature (59.7 percent), whereas some are time-lagged (7.5 percent) or longitudinal (32.8 percent articles).

The final criterion, the elimination of alternative explanations, suggests that if the influences of external forces are not eliminated it is impossible to assess the extent to which the external forces may be causing the effect (Van de Ven 2007). The ability to eliminate alternative explanations can be influenced by sampling design (Cook and Campbell 1979). Specifically, because of sampling bias, relationships reflected in the data may not accurately reflect the relationships in the population when small, nonrandom samples are used. Examining the sampling methods of the studies in our review, the majority (56.7 percent) examine differences between family and nonfamily firms. As such, the samples included both types of firms (391 family firms and 381 nonfamily firms, on average). The remainder of articles assesses family firm heterogeneity by examining agency theory within a sample of family firms only (498 family firms on average). Investigations use secondary data (64.2 percent) or primary data (35.8 percent). Of those using primary data, the average response rate is 23.80 percent. Samples of firms in North America (US and Canada) and Europe are equally represented (28 and 27 of the articles, respectively); firm samples from Asia (9 articles), Australia (2 articles), and South America (1 article) are also represented. Unfortunately, only one article used a random sample; in this study, the authors drew a random sample of 15,517 firms from the Italian population of nearly 5 million firms (Sciascia and Mazzola 2008).

Ruling out alternative explanations can also be accomplished by incorporating statistical control variables into the research design (Atinc et al. 2012; Becker 2005). By including theoretically meaningful control variables, variance in the dependent variable that is attributable to the alternative explanation is accounted for a priori, enabling researchers to assess the extent to which the remaining variance in the effect is explained by the hypothesized cause (Atinc et al. 2012). Given that regression is the data analysis approach in the overwhelming majority of investigations (60 of 67 articles; other approaches included structural equation modeling, cluster analysis, and canonical analysis), we expected a large percentage of articles in our data set to use control variables. Indeed, among the agency theory articles examined, all but one made use of control variables. In addition, authors often provided the proper rationale, theoretical reasoning, or citations to explain why a particular control variable was included. Schulze et al. (2003a, 2003b), for instance, argued that covariates were to reduce variance that was extraneous to the research question and could potentially confound the interpretation of the findings. In each article, the authors controlled for firm age, justifying this variable as it might be linked to the performance outcome studied due to a selection bias where older firms were in the sample just because they had survived and were successful.

Finally, by not accounting for endogeneity, scholars are not able to eliminate several alternative explanations, leading to incorrect generalizations. The problem of endogeneity arises when the cause that is hypothesized to have a particular effect is influenced directly by that effect. While a number of statistical techniques exist to test for such problems, these methods are just beginning to appear in the family business research (e.g., Patel and Cooper 2014). In fact, these methods were used in only 21 (31.3 percent) of the articles examined. This is particularly important in family firms where the involvement of the family clearly may influence the performance of the firm but the performance of the firm may also influence the family’s involvement, especially the family’s involvement in the day-to-day operations.

In summary, our review of the methodology employed in family firm agency theory studies reveals that our ability to effectively infer causality may be inhibited. Although roughly two-thirds of the studies include correlation matrices, only one-third are time-lagged or longitudinal, and only one-third rule out alternative explanations by accounting for endogeneity concerns. Furthermore, the samples are not random and often are relatively small with low response rates, which gives rise to generalizability concerns. Given the importance of research methods in establishing and validating causal inferences, we incorporate methodological remedies as we present our future research opportunities.