Family versus Non-Family Firm Franchisors: Behavioral and Performance Differences

Drawing from resource-based theory, we argue that family firm franchisors behave and perform differently compared to non-family firm franchisors. Our theorizing suggests that compared to a non-family firm franchisor, a family firm franchisor cultivates stronger relationships with franchisees and provides them with more training. Yet, we predict that a family firm franchisor achieves lower performance than a non-family firm franchisor. We argue, however, that this performance relationship reverses itself when family firm franchisors are older and larger. We test our hypotheses with a longitudinal dataset including a matched-pair sample of private U.S. family and non-family firm franchisors.


INTRODUCTION
terms of behaviors and performance in franchising as a key entrepreneurial activity (Ketchen, Short and Combs, 2011).
Scholars use various theoretical perspectives to study franchising (see Combs et al., 2011b;Combs, Ketchen and Short, 2011c); however, the resource-based view is key to efforts to explain the franchising phenomenon in general and franchising behaviors and performance in particular (Castrogiovanni, Combs and Justis, 2006a;Combs and Ketchen, 1999b;Combs et al., 2011b). Importantly, the resource-based view suggests that franchising behaviors and performance outcomes should differ considerably between family firms and non-family firms. This is because a main feature that distinguishes family firms from non-family firms is their distinctive bundles of resources that emerge through the interaction between the family and the business systems (Eddleston, Kellermanns and Sarathy, 2008;Habbershon, Williams and MacMillan, 2003;Sirmon and Hitt, 2003). Unique resource bundles, in turn, have important implications for franchising (Combs et al., 2011c), as sharing unique, yet complementary resources between the franchisor and the franchisee is a core element of franchising, leading to joint value-creating benefits (Combs and Ketchen, 1999b;Ketchen, Ireland and Snow, 2007).
Drawing from the resource-based view of the firm, we theorize that family firm franchisors behave and perform differently compared to non-family firm franchisors. In terms of behavior, we argue that a family firm franchisor establishes stronger relationships with franchisees and provides them with more training. We focus on these two facets of franchising as these are essential aspects of how to share resources and create value in franchising (Chirico et al., 2011a;Combs and Ketchen, 1999b;Ketchen et al., 2007). Strong relationships facilitate resource sharing (Chrisman, Chua and Kellermanns, 2009) and allow superior information exchange (Baucus, Baucus and Human, 1996;Dant and Nasr, 1998). As such, strong relationships can enable collaborating parties to form and use resources that have strategic value (Dyer and Singh, 1998). In addition, providing training to franchisees is a useful tool to familiarize them with the franchisor's procedures, operations, best practices, and management approaches (Gillis, Combs and Ketchen, 2014). As a result, training for franchisees constitutes an important learning opportunity where the trainees can develop and enhance their skills and capabilities (Miller, Le Breton-Miller and Scholnick, 2008). Given these expectations, extant franchising literature has investigated the strength of the franchisor-franchisee relationship (e.g., Chirico et al., 2011a;Davies, Lassar, Manolis, Prince and Winsor, 2011) and training activities (e.g., El Akremi, Perrigot and Piot-Lepetit, 2015; Gorovaia and Windsperger, 2013) to explain the franchising phenomena and related outcomes. With respect to performance, we theorize that family firm franchisors achieve a lower level of performance than non-family firm franchisors; however, we expect a reversal of this outcome when family firm franchisors are older and larger. Empirically, we rely on secondary data from FRANdata, which is a reliable source of objective information regarding franchise operations in the United States. Information is extracted directly from federal Franchise Disclosure Documents (FDDs). Our analysis of a longitudinal matched-pair sample of U.S. family-and non-family firm franchisors generally confirms our theoretical expectations.
Our results yield several contributions. First, we contribute to the corporate entrepreneurship literature. On a general level, we demonstrate that franchising is a relevant and unique type of corporate entrepreneurship (see Ketchen et al., 2011;Ucbasaran, Westhead and Wright, 2001) in the form of external corporate venturing (Sharma and Chrisman, 1999).
On a more specific level, we advance the literature concerned with corporate entrepreneurship in family firms (see McKelvie, McKenny, Lumpkin and Short, 2014). By focusing on family firms as a specific context (Davidsson, Low and Wright, 2001), we enhance our understanding of differences in entrepreneurial behavior between family and non-family firm franchisors and shed light on key contingency factors that affect performance. As such, we enrich the discussion about whether family firms are more or less entrepreneurial than non-family firms (Eddleston, Kellermanns and Zellweger, 2012;Randolph, Li and Daspit, 2017) and how this relates to performance differences (Miller and LeBreton-Miller, 2006;Villalonga and Amit, 2006). In addition, we advance our understanding of whether (or not) the family firm's commitment to entrepreneurial actions, such as franchising, erodes over time and with advances in firm size (Barringer and Bluedorn, 1999;Naldi, Nordqvist, Sjöberg and Wiklund, 2007).
Second, we contribute to the small but expanding body of research concerned with franchising within family firms. Drawing from resource-based arguments, we develop novel theorizing to predict the behavioral and performance-related implications of being a family firm franchisor in a franchising context. This allows us to better understand the franchising phenomenon and to isolate the related theoretical reasons for differences in franchising behaviors and performance among franchisors. Additionally, our theorizing facilitates the field's efforts to understand previous contradicting theoretical arguments and results regarding the effects of firm age and size on franchisor outcomes (e.g., Castrogiovanni, Combs and Justis, 2006b;Castrogiovanni, Justis and Julian, 1993;Combs and Ketchen, 2003).

Franchising as a Form of Corporate Entrepreneurship
Generally, franchising is understood as a long-term business arrangement wherein one firm (the franchisor) grants the right to market goods or services under its brand name and to use its business practices, processes, and routines to another firm (the franchisee) (Combs et al., 2011b;Combs et al., 2004;Combs, Michael and Castrogiovanni, 2009). The franchisee, in turn, leverages its idiosyncratic knowledge of the local environment resource, including competitors' activities, and applies it to the franchisor's business model (Combs et al., 2011b). Importantly, franchising represents a form of firm-level corporate entrepreneurship (Ketchen et al., 2011;Sharma and Chrisman, 1999). Sharma and Chrisman (1999: 20) define corporate entrepreneurship as "the process whereby an individual or a group of individuals, in association Electronic copy available at: https://ssrn.com/abstract=3532121 with an existing organization, create a new organization or instigate renewal or innovation within that organization" and suggest that a key element of corporate entrepreneurship is "external corporate venturing," where an existing organization creates semi-autonomous or autonomous entities outside its own boundaries that may vary in their degree of separateness from the parent company (Ellis and Taylor, 1987;Sharma and Chrisman, 1999;von Hippel, 1977). In the context of our work, the franchisor is the existing parent company that creates new organizational entities outside its own boundaries (i.e., the different franchisees).
Overall, franchising plays a key role in developed economies. For instance, in the United States, 745,290 (+1.6% compared to the previous year) franchise establishments in 2017 employed 7,881,000 (+3.1%) individuals with $713.2 (+5.6%) billion in output and contributed $425.5 billion (+ 5.1%) to the nation's gross domestic product (GDP). 1 These data suggest the importance of franchising and with respect to our arguments, the importance of franchising as a form of corporate entrepreneurship.

Franchising and Resource-Based Theory
Scholars use an array of theoretical perspectives, and combinations of them, to examine franchising. Examples of theories franchising scholars use include agency theory and resource scarcity considerations (e.g., Carney and Gedajlovic, 1991;Castrogiovanni et al., 2006b;Combs and Ketchen, 1999a). Explaining the emergence of franchising as a means of organizing (see Combs et al., 2011b;Gillis and Castrogiovanni, 2012) and describing key characteristics of the organizing process, such as the choice between company-owned and franchised units and the related ownership redirection hypothesis, are examples of issues explored by scholars using these two theories (e.g., Carney and Gedajlovic, 1991;Castrogiovanni et al., 2006b;Combs and Ketchen, 1999a). Scholars also use property rights theory to explain ownership redirection in franchising (Windsperger and Dant, 2006) and the emergence of multi-unit franchising (Hussain and Windsperger, 2013). Institutional theory (Barthélemy, 2011;Combs et al., 2009), signaling theory (Dant and Kaufmann, 2003), social exchange theory (Meek, Davis-Sramek, Baucus and Germain, 2011) or a combination of agency theory and institutional theory (Barthélemy, 2008) are also used frequently in franchising research (for an overview, see Combs et al., 2011b). However, multiple reasons lead us to draw from resource-based theory as our primary theoretical lens.
In our study, we seek to explain differences in franchising behaviors and performance between family and non-family franchisors. Resource-based theory is particularly appropriate to frame arguments around our research question. For instance, Combs and Ketchen (1999a: 204) argue that the resource-based view "with its emphasis on the link between organizational capabilities and competitive advantage seems likely to offer additional insight into franchising behavior." Castrogiovanni et al. (2006a: 40) suggest that "researchers should consider resourcebased theory as a complementary explanation of franchising behavior." With respect to firm performance, resource-based theory seeks to explain sustainable competitive advantage (Barney, 1991). To exploit a competitive advantage, "firms must possess resources that can be used to create inimitable and rare value for customers" (Ireland, Hitt and Vaidyanath, 2002: 428). Through different mechanisms, including training and strong relationships, franchisors and franchisees seek to find ways to accumulate resources internally that facilitate their efforts to create inimitable and rare value for customers.
Due to the relationship between resource-based arguments and the nature of franchising, many scholars use this theory to study franchising behaviors and resulting performance outcomes (e.g., Combs and Ketchen, 1999b;Gorovaia and Windsperger, 2013;Wu, 2015).
Moreover, resource-based theory helps predict franchising behaviors and performance in that it notes that firms use their unique resource bundles as the foundation for value creation by Electronic copy available at: https://ssrn.com/abstract=3532121 developing and exploiting competitive advantages (Barney, 1991;Harrison, Hitt, Hoskisson and Ireland, 2001;Sirmon, Hitt and Ireland, 2007). Value-creating resources are assets and capabilities that are valuable, rare, inimitable, and non-substitutable (Combs, Ketchen, Ireland and Webb, 2011a;Morrow, Sirmon, Hitt and Holcomb, 2007). Sirmon et al. (2007: 273) highlight this point by noting that "…the RBV suggests that possessing valuable and rare resources provides the basis for value creation." A main motive for using franchising as an organizational form is to structure, bundle and leverage (i.e., resource management; Sirmon et al., 2007) scarce resources such as capital and managerial resources between the franchisor and franchisees to foster rapid growth and performance improvements (Combs and Ketchen, 1999b;Oxenfeldt and Kelly, 1969).
Resource-based theory, and in particular the 'extended' resource-based logic, also predicts that firm-specific internal resources are not the only source of competitive advantages; rather, organizations can benefit from a wide array of external resources without having full control over them (see, for instance, Dyer and Singh, 1998;Gulati, Nohria and Zaheer, 2000;Ireland et al., 2002). That is, organizations can engage in interfirm cooperation arrangements to share resources across organizational boundaries and generate joint value-creating benefits (Cao, Simsek and Zhang, 2010;Lavie, 2006;Li, Eden, Hitt and Ireland, 2008). Given that a key aspect of franchising is interfirm resource sharing between franchisor and franchisees, this logic applies particularly well to the franchising context (Combs et al., 2004).
In a franchisor-franchisee relationship, sharing complementary resources is foundational to efforts to mitigate resource constraints and enhance the value-creating ability of both parties (Combs and Ketchen, 1999b;Ketchen et al., 2007). Put differently, it is difficult for a single firm to possess or control all resources required to outperform rivals. Through franchising, the franchisor and franchisee can combine their resources to create bundles with the capability to create more value than either party would create acting independently Electronic copy available at: https://ssrn.com/abstract=3532121 (Harrison, Hitt, Hoskisson and Ireland, 1991;Makri, Hitt and Lane, 2010). Through competitive advantages developed by integrating unique resource bundles, the franchisor and franchisee increase the likelihood of reaching desired performance objectives (Combs and Ketchen, 1999b). Experiences resulting from firm age and additional assets resulting from firm size enhance a firm's ability to recognize and use resources it would like to add to its resource portfolio as a foundation for creating additional value (Castrogiovanni et al., 2006b).
As noted previously, we believe that resource-based theory is particularly appropriate to investigate behavioral and performance differences between family and non-family firm franchisors. This is because a crucial characteristic of family-owned firms that distinguishes them from non-family firms is the ability to create unique resource bundles that are a product of idiosyncratic interactions between the family and the business. Some of these unique interactions may occur, for instance, during training sessions franchisors provide to franchisees.
Compared to non-family firms, family firms are in fact more complex, dynamic, and resource rich, particularly in terms of the set of intangible resources (Eddleston et al., 2008;Sirmon and Hitt, 2003) that originate from interactions among the family, its individual members, and the business (Habbershon et al., 2003;Habbershon and Williams, 1999). As such, these unique resource bundles provide them with a resource base for which non-family firms find duplication difficult (Pearson, Carr and Shaw, 2008) and which has important implications in a franchising setting (Chirico et al., 2011a;Combs et al., 2011c). Examples of resources a franchisor may provide are copyrights, patents on formulas, registered brands, networks, procedures and operations, best practices, purchasing power, and management/marketing assistance and advice (Gillis et al., 2014). Resources shared by a franchisee may be financial capital, local knowledge (e.g., in relation to the market, the culture, customer preferences and needs), local networks (i.e., alliances with other franchisees, marketing connections), and local reputation and goodwill (e.g., charitable giving to small league teams) (Chirico et al., 2011a;Combs et al., 2004). By sharing complementary resources such as these, benefits accrue to the franchising parties. In addition, an expectation is that both parties will experience greater economies of scale in organizational functions such as purchasing, marketing, and legal affairs. For instance, some franchising studies examine practices through which social relationships and training facilitate efforts to retain valued human capital (e.g., El Akremi et al., 2015;Perdreau, Le Nadant and Cliquet, 2015;Stanworth, Stanworth, Watson, Purdy and Healeas, 2004).
Reputation, defined as the general level of favorability stakeholders have of a particular firm (Deephouse and Jaskiewicz, 2013), is another valuable resource for franchising parties.
Describing this value, Flanagan and O'Shaughnessy (2005: 445) note that reputation is "one of the most important strategic resources" for a firm to possess. Because of its value, reputation is often an important contributor to a firm's efforts to create sustainable competitive advantages (Griessmair, Hussain and Windsperger, 2014).
In the franchising context, evidence suggests that a franchisor's positive brand name reputation has the potential to enhance its performance (see Barthélemy, 2008;Combs and Ketchen, 1999b;Wu, 2015). Identity overlaps between the family and its business (Dyer and Whetten, 2006;Zellweger, Nason, Nordqvist and Brush, 2013) explain some of the reputation's importance as a resource in a family business. This overlap implies a family's explicit concern about the firm's reputation with external stakeholders as it can in turn affect the reputation of the family as well. The evidence suggests that reputation is a unique family-influenced resource with numerous benefits at the firm level (Habbershon et al., 2003;Pearson et al., 2008) such as enhanced entrepreneurial behavior (Clinton, Nason and Sieger, 2013;Sieger, Zellweger, Nason and Clinton, 2011) or improved performance and value creation (Rindova, Williamson and Petkova, 2010;Roberts and Dowling, 2002;Shane and Cable, 2002). Due to its importance, we consider reputation as a critical resource for franchisors in our study. 2

Franchising and the Family Firm
There are many definitions of family firms (Sharma, 2004); however, family involvement in terms of ownership and management are the most commonly used criteria to identify a family business (see, for instance, Astrachan, Klein and Smyrnios, 2002). This position is highlighted by Arregle, Hitt, Sirmon and Very (2007: 87) who state that "a business firm may be considered a family business to the extent that its ownership and management are concentrated within a family unit." As such, family firms are organizations that a family owns and manages, as depicted commonly in many empirical family business studies (e.g., Chirico et al., 2019;Miller et al., 2007).
Evidence suggests a strong presence of family firms in the franchising context (e.g., In our theorizing, using resource-based arguments, we focus on two of the most crucial elements of a franchising agreement: the relationship between a franchisor and its franchisee and the training the franchisor provides to the latter (see, for instance, Combs et al., 2004). As discussed earlier, these aspects are essential for resource sharing as a foundation for creating value through franchising (Chirico et al., 2011a;Combs and Ketchen, 1999b;Ketchen et al., 2007). Because of their importance, scholars investigate these two franchising aspects frequently (e.g., Davies et al., 2011;El Akremi et al., 2015;Gorovaia and Windsperger, 2013).
Electronic copy available at: https://ssrn.com/abstract=3532121 In our study, we also consider performance outcomes and related important contingency factors (Combs and Ketchen, 2003).

HYPOTHESES The Franchisor-Franchisee Relationship
We suggest that family firm franchisors are more likely to build stronger relationships with their franchisees than non-family franchisors. There are several reasons for this expectation.
First, a long-term perspective and time horizon is a feature attributed commonly to family firms (Zellweger, Kellermanns, Chrisman and Chua, 2012). In turn, achieving success over time requires family firm franchisors to establish strong, long-lasting interfirm relationships with their franchisees (see Miller et al., 2007). As such, compared to non-family franchisors, who Owing to their commitment to sharing complementary resources as a means to accumulate value-creating assets internally (see Sirmon et al., 2007: 279) for the benefit of both parties in the franchising system and to be able to facilitate long-term success in the process of doing so (Chirico et al., 2011b), family firm franchisors are likely to invest more time and effort in building and cultivating the franchisee-franchisor relationship than a non-family firm franchisor. Drawing from a generational perspective and the view of patient capital as an important resource for long-term success, Sirmon and Hitt (2003) suggest that family firms are in a position to establish richer interfirm relationships. In turn, these relationships facilitate efforts by the family firm franchisor to create value when collaborating with its franchisees.
Relatedly, to ensure long-term success of its franchising business, family firm franchisors may also be more committed to expending extra efforts and resources to build stronger relationships with franchisees. Using resources in this manner likely generates patient capital and survivability capital -which are often a product of family members combining business and family resources (Habbershon et al., 2003) -for the franchising relationship (see Eddleston and Kellermanns, 2007;James, 1999;Zellweger, 2007).
Social capital is a resource with the potential to facilitate firms' efforts to create joint values (Dyer and Singh, 1998;Ireland et al., 2002;Ketchen et al., 2007). As a resource, social capital involves trust, exchange, generosity, and solidarity (Bubolz, 2001). Due to close ties and commonly strong commitments to achieving objectives, a family firm is an ideal environment in which to establish social capital through which the firm creates value for stakeholders (Coleman, 1988). Thus, a family firm's unique social capital is another resource distinguishing family firm franchisors from non-family franchisors. Arregle et al. (2007: 77) speak to this issue, noting that the social capital of a family firm is "probably one of the most enduring and powerful forms of social capital." In the franchising context then, compared to a non-family firm franchisor, a family firm franchisor is likely to better use its rich social capital to build and sustain stronger and more stable relationships with its franchisees with accepted norms and routine interactions as the foundation. This facilitates interfirm resource sharing (Chrisman et al., 2009;Duschek, 2004;Lavie, 2006) and information exchange (Baucus et al., 1996;Dant and Nasr, 1998;Ketchen et al., 2007) and ultimately enhances the likelihood of the franchise's long-term continuity.
Lastly, because of identity overlaps between a family and its firm (Dyer and Whetten, 2006;Zellweger et al., 2013), the family firm franchisor has strong incentives to establish and ensure a favorable franchisor reputation with external stakeholders. A strong and stable relationship between the family firm franchisor and its franchisees contributes to such a positive reputation. In contrast, a weak relationship between these parties has the potential to affect the reputation of the family firm franchisor and the family itself negatively. In sum, we predict the following: Hypothesis 1: Compared to a non-family firm franchisor, a family firm franchisor builds stronger relationships with its franchisees.

Franchisees' Training
A core part of the franchising agreement is the training of franchisees (Combs et al., 2004) through which resources are shared and competencies built beyond organizational boundaries (Das and Teng, 2000;Ireland et al., 2002;Lavie, 2006). Building competencies supports efforts by the franchising partners to respond to changing customer needs consistently and effectively over time.
A key objective franchisors have in training franchisees is to develop the human capital embedded within those units . Sharing and providing access to procedures, operations and routines, and general management/marketing knowledge and advice to franchisees is one path through which franchisors seek to develop human capital in those units (Gillis et al., 2014). Classroom training and/or on-the-job training are actions franchisors take to develop human capital in their franchisees (see El Akremi et al., 2015;Perdreau et al., 2015;Stanworth et al., 2004).
For several reasons, we expect that family firm franchisors are likely to provide more continuous training to their franchisees than non-family firm franchisors. On a general level, evidence indicates that family firms provide an environment that fosters a family-oriented workplace, which in turn inspires greater employee commitment and loyalty (Gedajlovic et al., 2012;Habbershon and Williams, 1999). Put differently, employers and employees build trust and even friendships, which in turn results in more motivated and loyal employees.
Consequently, the actions of these individuals and groups have positive effects on the family firm's prosperity and long-term continuity (Ling and Kellermanns, 2010).
Therefore, family firms commit considerable resources to ongoing training and learning opportunities, including internships that develop employees more fully as a valuable resource and provide critical motivational opportunities for personal growth (Miller et al., 2008).
Compared to non-family firms, family firms are more likely to train and further develop their employees on an ongoing basis because of their strong commitment to family firm continuity (Miller et al., 2008). We argue that these patterns apply to the franchisor-franchisee context as well. This means that compared to a non-family firm franchisor, a family firm franchisor is more likely to devote time and resources to provide training to franchisees. This additional training, both in class and on-the-job, has the capability to expand the skills of each franchisee's human capital and to develop social capital in each franchisee unit by developing feelings of belonging to the business and the family (Miller et al., 2008). Such behaviors may further support a friendly and an informal organizational culture based on loyal and cohesive group relationships (Zahra, Hayton and Salvato, 2004). These relationships, in turn, provide motivation for the family firm franchisor and its franchisees to work collaboratively to achieve common interests.
Family firm reputation is another key reason we anticipate family-firm franchisors to provide more training to franchisees than will non-family firm franchisees. As mentioned previously, a family firm's reputation is a unique family-influenced resource (Habbershon et al., 2003;Pearson et al., 2008;Sieger et al., 2011). Because of identity overlaps between the family and the business (Dyer and Whetten, 2006;Zellweger et al., 2013), family members express concerns about the firm's reputation with external stakeholders in that it affects their family reputation as well. Within the franchising context, this means that when a franchisee exhibits deviant behaviors such as not meeting the requirements and not following the guidelines and procedures as defined in the franchising agreement, this likely will have a negative effect on the reputation of the franchisee, the franchisor, and ultimately, the family owning the franchise (Sieger et al., 2011). Therefore, compared to non-family firm franchisors, family firm franchisors have a stronger incentive to ensure that franchisees behave as they should; effective ongoing training programs (Miller et al., 2008) are a path with the potential to reduce the likelihood of inappropriate and ineffective franchisee behaviors. Put differently, because of the desire to avoid negative reputation implications for the family through a franchisee's behaviors, we expect family firm franchisors to invest more in franchisee training than non-family firm franchisors.
Finally, as mentioned above, family firms typically have a longer-term perspective and time horizon than non-family firms . We suggest that this perspective also manifests itself through training that seeks to provide franchisees with important managerial and operational input, ensuring that franchisor and franchisee "remain on the same page" in the long run. Following the above arguments, we therefore posit: Hypothesis 2: Compared to a non-family firm franchisor, a family firm franchisor provides more training to its franchisees.

Franchisor's Performance
Resource-based logic suggests that family firm franchisors are more willing and able to seek to accumulate resources by sharing complementary resources with franchisees (e.g., human capital through training). In turn, we expect that the sharing of resources will yield performance capabilities that are specific to each franchising situation because of the unique resources themselves as well as the idiosyncratic interactions between a franchisor and its franchisees through which resources are used (Habbershon et al., 2003;Pearson et al., 2008;Zahra, 2010).
As a result, intensive sharing of unique resources across organizations has the potential to be a source of competitive advantage (Duschek, 2004;Lavie, 2006) and subsequently, of superior financial performance for family firm franchisors as compared to non-family firm franchisors.
We theorize, however, that this is not necessarily the case due to path-dependent behaviors that influence family firm franchisors' financial outcomes (see Zahra, 2005). In fact, an option family firms often use is retention of the status quo. Doing this mires family firms "in a single way of seeing and doing things," thus "converting a formula for success into a path toward failure" (Miller, 1993: 116, 122) while supposedly preserving the family and the business. Specifically, as Sirmon et al. (2007) explain, path dependency can limit the design of future resource-leveraging strategies. This is important in that evidence suggests that family firms often engage in path-dependent behaviors (König, Kammerlander and Enders, 2013) that provide a sense of familiarity for family members who perceive past solutions as being less risky than attempting a de novo solution to exploiting an opportunity (Hoskisson, Chirico, Zyung and Gambeta, 2017;Miller, Steier and Le Breton-Miller, 2003;Zahra, 2005). Such path dependence may increase the risk that the family firm franchisor and related franchising activity could fall into what Ahuja and Lampert (2001) call a familiarity trap that "limit[s] the openness to information and to alternative ways of doing things, producing collective blindness" (Nahapiet and Ghoshal, 1998: 245).
Thus, family firm franchisors may tend to replicate inherited organizational routines and strategic perspectives, especially when the resources in question contributed to prior success (Sirmon and Hitt, 2003). Subsequently, the emphasis may shift more to addressing problems and making decisions in light of past behaviors rather than seeking to identify and exploit new opportunities (Chirico et al., 2011a). The organizational rigidity resulting from this type of path dependence may prevent the franchising business from developing an ability to adapt to changing circumstances. When this happens, family-firm decision makers may envision conditions surfacing in the external context as threats rather than as potential opportunities.
Such a perspective reinforces a cycle of defensive behaviors that, at least in part, decision makers use to preserve the status quo (Miller et al., 2003).
Moreover, while we expect family firm franchisors to establish stronger relationships with and to provide more training for their franchisees by sharing complementary resources, this expectation does not necessarily imply that family firm franchisors will achieve higher performance than non-family firm franchisors. This can happen as family firm franchisors and their franchisees may share interfirm resources that are obsolete or competitively inferior. In turn, this type of sharing finds firms using outdated behavioral paths. Typically, for instance, the training programs and routines the franchisor provides to the franchisee reside in the franchisor's established routines-routines that at some point support the status quo (Hackman, Brousseau and Weiss, 1976). Put differently, as an organizational resource, long-established routines in family firm franchisors may find the firms falling into a familiarity trap (Ahuja and Lampert, 2001). As organizations gain increasing familiarity with their resources and bundles of them and the outcomes that are possible by using them, the tendency to rely too much on them increases. A negative unintended consequence of the familiarity trap is that a firm's ability to identify new resource combinations as a means of improving performance decreases (Ahuja and Lampert, 2001). In addition, the strength of the desire to maintain existing family relationships (Kellermanns and Eddleston, 2004) and the routines making them possible may result in family firm franchisors failing, for instance, to upgrade the quality of their training programs for franchisees on a consistent basis. Without appropriate enhancements to the programs' contents, franchisees' performances may decline.
Owing to the intention to avoid conflicts and preserve harmony while sharing resources, family firm franchisors and their franchisee may also tend to agree on the "smallest common denominator" to avoid placing their relationship at risk (Hoskisson et al., 2017;Kellermanns and Eddleston, 2004). This tendency though may result in less innovative and entrepreneurial behavior, outcomes that likely fail to enhance the franchising arrangement's performance, particularly from a financial perspective. This reasoning leads us to hypothesize that a family firm franchisor achieves lower performance than a non-family firm franchisor. Stated formally, we posit that: Hypothesis 3: Compared to a non-family firm franchisor, a family firm franchisor achieves lower performance.

Franchisor's Age and Size
We theorized that family and non-family firm franchisors differ in terms of performance.
Following previous studies (Castrogiovanni et al., 2006a;Combs and Ketchen, 2003;Combs et al., 2004), we now introduce the franchisor's age and size as related crucial contingency factors.
Firm age and size are fundamental elements of dyadic relational models (e.g., Bordonaba-Juste and Polo-Redondo, 2008;Castrogiovanni et al., 2006a;Weaven and Frazer, 2003). Evidence suggests that these variables are important to efforts to explain firm survival and performance (see, for instance, Bruderl and Schussler, 1990;Ling, Zhao and Baron, 2007;Stinchcombe, 1965). Additionally, some scholars study the effects of firm age and size on franchisor outcomes. Results from these studies show mixed effects (e.g., Combs and Ketchen, 2003;Kosová and Lafontaine, 2010;Lafontaine and Shaw, 1999). Some results indicate that franchisor age and size are reasonable proxies for learning outcomes (e.g., Castrogiovanni et al., 2006a). In other cases, results suggest that instead, they are associated with unit obsolescence (e.g., Castrogiovanni et al., 1993).
Although these studies yield mixed results, we assert that both learning outcomes and unit obsolescence may apply in the franchising context. However, we believe that their respective occurrence is more or less likely to happen depending on whether the franchisor is a family or a non-family firm. The reason for this expectation is that a firm's governance structure directs resource allocation and utilization decisions (Makadok, 2001;Makadok, 2003); accordingly, governance characteristics such as family versus non-family firm status likely affect franchising-based performance gains (Combs et al., 2004). For the reasons outlined below, we argue that increased firm age and size of the franchisor combine to reverse the performance disadvantage of family firm franchisors, with family firm franchisors then outperforming non-family firm franchisors.
First, we theorize that franchisor age and size are more likely to lead to positive performance outcomes for family firm franchisors than for non-family firm franchisors. Family firms are regarded as high-trust and non-opportunistic organizations "governed by underlying informal agreements based on affect rather than on utilitarian logic or contractual obligations" (Gomez-Mejia, Nunez-Nickel and Gutierrez, 2001: 82). These family firm characteristics may promote behavior that enables family firm franchisors to perform best as they age and grow, thus offsetting path-dependency and the related performance-constraining behaviors of family firm franchisors we proposed previously. For example, family firm franchisors' focus on the long-term success of the franchising activity may lead to effective and efficient use of existing resources and to important financial benefits over time (Strike, Berrone, Sapp and Congiu, 2015;Zellweger and Astrachan, 2008). Additionally, large family firm franchisors may be particularly skilled in managing relationships and sharing resources with an increasing number of franchisees and unit managers (Udell, 1973), thus contributing to positive outcomes. As such, when a family firm franchisor is older and larger, important family firm-specific learning processes (Castrogiovanni et al., 2006a) occur, strengthening the value-creating potential of the resources shared with franchisees; ultimately, we expect this to have positive performance implications (Duschek, 2004;Lavie, 2006). Relatedly, the family firm franchisor's long-term perspective, care and devotion to franchisees and the franchising system of activities, along with the use of patient capital and survivability capital (Sirmon and Hitt, 2003), have the potential to mitigate path-dependent behaviors and thus potentially enhance performance outcomes as the family firm ages and grows.
Second, some scholars argue that franchisor age and size are associated with unit obsolescence (e.g., Castrogiovanni et al., 1993). Yet, we believe that the danger of unit obsolescence is not as prominent for family firm franchisors compared to non-family firm franchisors. This is because compared to family firm franchisors (Chirico et al., 2011a), nonfamily franchisors appear to be less capable of sharing resources and managing long-term relationships with an increased number of franchisees and unit managers (Bordonaba-Juste and Polo-Redondo, 2008;Bruderl and Schussler, 1990;Castrogiovanni et al., 1993). More specifically, research suggests that overall, mature and large non-family franchisors find it increasingly difficult to maintain or strengthen their relationships and resource sharing with both franchisees and their managers over time, thus making it more challenging for franchisors to achieve positive financial outcomes (Bordonaba-Juste and Polo-Redondo, 2008; Watson and Johnson, 2010;Weaven and Frazer, 2003). Supporting this view, Bordonaba-Juste, Lucia-Palacios and Polo-Redondo (2011) and Castrogiovanni et al. (1993) report that franchisor age has a positive effect on failure while Kosová and Lafontaine (2010) report a negative effect of franchisor age and size on chain growth. When the non-family franchisor is older and larger, franchisees appear to have less confidence that they are receiving significant value from the franchisor (Brizek, 2004). Such mature and large franchisors tend to become multi-unit operators (Bradach, 1997;Weaven and Frazer, 2003) to reduce interactions with both franchisees and their managers, thereby mitigating potential conflicts with them (e.g., Kaufmann and Kim, 1995;Winter, Szulanski, Ringov and Jensen, 2012).
Additionally, evidence shows that firm age and size are antecedents of a favorable reputation, also in the franchising context. Referring to age, corporate reputation in general is socially constructed (Lange, Lee and Dai, 2011), accumulates slowly (Fombrun, 1996), and is based on the multiple actions associated with an organization's past (Flanagan and O'Shaughnessy, 2005). As such, companies in business over a longer time period can be assumed to have satisfied their stakeholders (at least minimally) over the years (Delgado-García, De Quevedo-Puente and De La Fuente-Sabaté, 2010;von Weizsacker, 1980), which contributes to the building of a positive reputation (Kosová and Lafontaine, 2010). With respect to firm size, Fombrun and Shanley (1990) found a strong positive link between this organizational characteristic and corporate reputation. A possible explanation is that firm size signals a firm's access to resources and enhances visibility. An implication of visibility is that different stakeholders are aware of the firm's actions and can closely monitor its behavior. This, in turn, leads to lower expropriation of firm value and, ultimately, a more desirable or enhanced corporate reputation (Delgado-García et al., 2010). A number of studies confirm the positive link between size and reputation (see Lange et al., 2011;Staw and Epstein, 2000). Therefore, the evidence suggests that the older and larger a franchisor is, the better should be its reputation.
A desirable reputational resource has the potential to contribute positively to forming and exploiting a competitive advantage, ultimately resulting in enhanced firm performance (Rindova et al., 2010;Roberts and Dowling, 2002;Shamsie, 2003).
Thus, we conclude that reputation becomes more important in predicting franchisor performance with increases in franchisor age and size. This is important and intriguing in that on a general level, research results suggest that family firms tend to care more about their firm's reputation than do non-family firms (Berrone, Cruz, Gomez-Mejia and Larraza-Kintana, 2010; Deephouse and Jaskiewicz, 2013). As discussed earlier, because of the business and family identities' overlap, family firm members seek to ensure that their firm has a positive reputation with stakeholders (including customers and suppliers) in that a negative reputation for the firm with external stakeholders would also yield a negative reputation for the family (Dyer and Whetten, 2006;Zellweger et al., 2013). Therefore, compared to non-family firm franchisors, family firm franchisors have greater incentives to consider their firm's reputation as a vital resource requiring attention (Habbershon et al., 2003;Pearson et al., 2008;Sieger et al., 2011).
Building on these arguments, we suggest that family firm franchisors, compared to non-family firm franchisors, have a stronger commitment to establishing a favorable reputation; in turn, a favorable or positive reputation increases its value-creating potential with increases in the franchisor's age and size. That is, when a family firm franchisor ages and increases in size, reputation gains in relevance as a predictor of firm performance. Taken together, this leads us to suggest that because of reputation-related dynamics, family firm franchisors outperform nonfamily firm franchisors when a franchisor is older and larger.
As a whole, there are reasons to expect that the performance pattern suggested in Hypothesis 3 will reverse in direction with increasing age and size. Formally stated: Hypothesis 4: A three-way interaction among franchisor age, firm size, and family firm status positively affects performance in such a way that compared to a nonfamily firm franchisor, a family firm franchisor achieves higher performance when the franchisor is older and larger.

Sample
Data regarding franchisors and franchisees are difficult to collect. For this study, we used a dataset from FRANdata (Information for the Franchise World) to test our hypotheses. Founded in 1989, FRANdata is the franchise industry's main source for objective information and analyses in the United States. Using its library of FDDs along with both primary and public sources, FRANdata supports franchisors' research and competitive intelligence functions, helps franchisees evaluate different concepts, provides information and analyses for legal and financial organizations, and provides marketing access to franchisors and franchisees.

FRANdata is the exclusive contractor for the Small Business Administration Franchise Registry
and is the only comprehensive source for FDDs outside of the U.S. federal government. 3 We tested our hypotheses using a longitudinal sample of private U.S. family and nonfamily firm franchisors included in the FRANdata source using an exact matched-pair design (Kennedy, 2008). We took three primary steps to construct the sample. First, FRANdata identified family firm franchisors by manually inspecting a randomly selected set of franchisors included in that group's database. FRANdata identified one hundred firms as family firm franchisors if two or more executives had the same family name; this is a well-established approach in empirical family business research (see, for instance, Deephouse and Jaskiewicz, 2013;Gomez-Mejia et al., 2001;Miller, Minichilli and Corbetta, 2013). Second, for each of these family firm franchisors, FRANdata identified a matching non-family firm franchisor through the following matching variables: industry, subsector, and firm size, 4 leading to a sample of 200 companies. Third, we ensured that the firms identified as family firms by FRANdata are owned and managed by a family, meaning that ownership and management are concentrated within a family unit (Arregle et al., 2007;Astrachan et al., 2002). We did this by checking company websites, franchise industry reports, and newspaper articles, with supportive results. In addition, we called all franchisors to confirm the family firm status. As some franchisors have a very restrictive policy of answering questions, we were not able to talk to all parties we attempted to contact. Even in these cases though, we were able to identify a sufficient amount of secondary material as supportive documentation that firms were family-owned and managed. In the end, this procedure led us to exclude only one company that we could not identify as either a family or a non-family firm. 5 All these franchisors have their home office location in the US, and the franchisees reported in the FDDs are US-based as well. We compared the mean values of age, size, and industry of this set of firms with those of all other firms included in the FRANdata database with t-tests and chi-square tests. We found no significant differences, a result indicating that our sample is representative of the initial population. We retrieved data including the years from 2003 to 2007; thus, we do not cover the global economic crisis that started in 2008 which affected the behavior and performance of family and non-family firms in the following years considerably -comprising those active in the franchising sector (IHS Global Insight, 2011). 6 Therefore, our data collection period has a positive effect on the generalizability of our findings. 7

Measures
Dependent Variables. Franchisor-franchisee relationship. We relied on two variables as a proxy for the strength of the franchisor-franchisee relationship with the first one being the number of contracts canceled by the franchisor or franchisees (Klein, 1995;Spinelli and Birley, 1996). Breach of contract provisions, failure to make royalty payments, non-adherence to quality standards, and failing to meet sales goals are examples of reasons for contract cancellations (Gandhi, 2014). 8 The second proxy is the number of actions per year brought to court (litigation up to 10 years old) by either the franchisor or franchisees to enforce a particular right connected with the franchising activity (see Antia, Zheng and Frazier, 2013;Winsor, Manolis, Kaufmann and Kashyap, 2012). As expected, these two measures are highly correlated (coeff.=.54; p<.001). Training for franchisees: We relied on the total number of training hours (in class and on-the-job training) the franchisor required the franchisees to attend, as indicated in the franchising contract (see Barron, Berger and Black, 1997;Konings and Vanormelingen, 2015). 9 Performance: We used an accounting-based measure to capture performance -the log of Return on Assets (ROA) (net income / total assets). ROA is an indicator of how profitable a 6 An example of a large franchisor going bankrupt is Mrs. Fields Famous Brands, the parent company of both the cookie chain "Mrs. Fields" and the frozen yogurt chain "TCBY." See https://www.cnbc.com/2014/10/31/10famed-american-franchises-that-faced-financial-ruin.html 7 We expect our results to hold in more recent datasets as well. Franchising is a very stable type of business in terms of arrangements between franchisors and franchisees. For instance, in 1979 the Federal Trade Commission passed the Franchise Disclosure Act (Rule 436.1), requiring 23 sections that a franchise disclosure document (FDD) is to contain. This Act was amended only in 2007 with very marginal changes, which indicates that the core practices and basic tenets associated with franchising have remained very stable over long periods of time (Judd and Justis, 2008). 8 Cancelled contracts do not include retirement. In the case of retirement, parties must still satisfy a contract's terms. FRANdata captures retirement in a different category. We thank an anonymous reviewer for this insight. 9 As such, the "provision of training" is measured by capturing the amount of "required training." We note that franchisors might also provide additional training to franchisees on a voluntary basis. This possibility, however, is not captured in our data given that the corresponding information is not available from FRANdata.
company is relative to its total assets and is a commonly used and well-established measure of performance (e.g., Barnett and Salomon, 2012;Karniouchina, Carson, Short and Ketchen, 2013;Zhang and Rajagopalan, 2010). In addition, ROA is particularly useful in the context of our research because it captures the degree to which franchisors are deploying firm resources effectively as a means of creating value. We also relied on alternative measures as proxies of performance as reported in the robustness test section.

Independent Variables.
We used a binary measure of family firm franchisors that was coded '1' if the franchisor is a family firm and '0' if the franchisor is a non-family firm, based on FRANdata's identification (two or more executives with the same family name) and our corresponding verification procedure (family ownership and managerial control). Additionally, to capture the variance within family firm franchisors, we used a continuous measure of family involvement indicating the number of family executives managing the firm (Deephouse and Jaskiewicz, 2013;Gomez-Mejia et al., 2001). All firms without family executives were considered non-family firm franchisors and were coded as 0, making the variable left censored (see Chrisman and Patel, 2012). We measure franchisor age as the number of years the firm had been operating as a franchisor. We measure franchisor size as the number of units owned and franchised by the franchisor. We logged both franchisor age and size because they were not normally distributed.
Covariates. In all our models, we controlled for franchisor age and size given that they may affect the behavior and performance of franchisors (Castrogiovanni et al., 2006b;Combs et al., 2009). In fact, franchisor age and size are regarded as standard control variables in franchising studies (see, among others, Barthélemy, 2008;Combs and Ketchen, 1999b;Sorenson and Sorenson, 2001). In addition, scholars highlight the importance of industry-level controls in franchising (Michael, 2003). Another important aspect commonly controlled for in the context of franchising behavior and performance is the distinction between franchised and Electronic copy available at: https://ssrn.com/abstract=3532121 company-owned units/outlets (Barthélemy, 2008;Combs and Ketchen, 1999b;Perdreau et al., 2015). Therefore, we controlled for the case in which a franchisor has company-owned units or not (coded "1" if yes, "0" otherwise). Additionally, while some studies use measures for competitor density (Litz and Stewart, 2000), market segment (Sorenson and Sorenson, 2001), or capital intensity (Madanoglu, Lee and Castrogiovanni, 2011), we used the more common sector variables to control for industry effects (El Akremi et al., 2015;Gorovaia and Windsperger, 2013). More specifically, we followed the SIC industry classification and used the construction and manufacturing industry as our reference category while employing three different dummy variables indicating whether a franchisor is active in retail and wholesale trade, finance, or services, respectively. To control for time dependency, we also incorporated the log of time into the analyses (that is, log (year of observation -first year of observation + 1), which is superior to using year dummies (see Box-Steffensmeier and Jones, 2004). Finally, to be conservative in our tests and to isolate the effects of interest as precisely and parsimoniously as possible, we controlled for the other dependent variables that we consider in our study in the respective models. For instance, in Models 1 (Tables II and III), we use the number of contracts canceled as the dependent variable; still, we added court actions, training, and performance as additional control variables.
Controlling for Endogeneity. It is possible that franchisors' relationships with franchisees, training for franchisees, and performance are endogenous to the unique features of the family firm franchisor versus the non-family firm franchisor. Stated somewhat differently, factors that might influence the relationship between franchisors and franchisees (such as contracts canceled and court actions), training, and performance could also influence the desirability of maintaining the franchisor firm as a family business.
We employed a two-stage residual inclusion (2SRI) model (see Patel, Criaco and Naldi, 2018;Terza, Basu and Rathouz, 2008) to control for potential endogeneity. The 2SRI estimator is similar to the linear two-stage least squares estimator except that in the second-stage regression, the first-stage predictors do not replace the endogenous variables; instead, first-stage residuals are included as additional regressors. We identified two instrumental variables: a) the location of the franchisor's headquarters and b) the number of years in business before starting the franchising activity to correct for potential endogeneity. Both instruments may affect a family's ability to extract private benefits from firm ownership and management while not directly affecting the franchisor-franchisees relationship, support, and performance (Bird and Wennberg, 2014;Miller et al., 2008). In the first stage, we used the instrumental variables to compute estimated values of the problematic predictors; we then used those computed values in the second stage to estimate a model of the dependent variables (Kennedy, 2008;Wooldridge, 2002). Thus, we controlled for the endogeneity score in all analyses (see Tables II-III; Chrisman and Patel, 2012).

RESULTS
We used panel data analysis with a random effect specification (Kennedy, 2008;Wooldridge, 2002) in Stata to test the hypotheses. In Table I, we present the descriptive statistics and correlations among the study's variables. We conditioned our statistical analyses on the matched cases and controls FRANdata identified and tested the hypotheses within different models, as reported in Tables II and III. In Table II, we report the estimated coefficients when we use the family firm dummy variable. In Table III Kennedy, 2008;Wooldridge, 2002).
Before creating the interaction terms, we standardized the variables to moderate multicollinearity problems, thus overcoming distortion of the main effects that could arise Electronic copy available at: https://ssrn.com/abstract=3532121 because of the tendency of main effects and interaction terms to be highly correlated (Aiken and West, 1991). Inspection of the variance inflation factors (VIFs) showed that multicollinearity was not a concern: all VIF coefficients were lower than 5 (Hair, Black, Babin, Anderson and Tatham, 2006).

------------------------------------------------------------------------INSERT TABLES I, II AND III ABOUT HERE ------------------------------------------------------------------------
Hypotheses 1 and 2 respectively predict that a family firm franchisor a) builds stronger relationships with franchisees and b) provides more training for them than a non-family firm franchisor does. The results in Models 1 (dependent variable: contracts canceled) and 2 (dependent variable: number of court actions) of Table II and III show mixed support for Hypothesis 1. As expected, the coefficients for the family firm dummy and the number of family executives are both negative and significant in Models 1. Thus, the results confirm that with a family firm franchisor, fewer contract cancellations occur between the franchisor and its franchisees. However, both coefficients are non-significant in Models 2 (Tables II and III (Tables II and III) show the results from testing Hypotheses 3 and 4. Hypothesis 3 states that a family firm franchisor achieves lower performance than a non-family firm franchisor. The coefficients for the family firm dummy and the number of family executives are negative and significant in Models 4 of Table II and III, respectively. These results support Hypothesis 3. Finally, to test Hypothesis 4, we followed Dawson and Richter (2006) and entered the two-way interaction terms in Models 5 (family firm measures * franchisor age; family firm measures * franchisor size; franchisor age * franchisor size) and the three-way interaction term in Models 6 (family firm measures * franchisor age * franchisor size). As predicted, the three way-interactions for both the family firm dummy and the number of family executives were positive and significant (Models 6 in Table II and III, respectively). The plot of the interaction (Figure 1) confirms that old and large family firm franchisors achieve the highest level of performance. This evidence supports Hypothesis 4.
Third, to obtain a more fine-grained understanding of Hypothesis 4, we separated the total company units into company-franchised units and company-owned units -an important distinction that is subject to very intensive discussions in the literature (Castrogiovanni et al., 2006b). Specifically, we ran the three-way interaction effects separately for companyfranchised and company-owned units (see Models 1 to 4 in Table IV) while removing the control variable indicating whether a franchisor has company-owned units or not. Interestingly, both interactions (family firm measures * firm age * franchised units; family firm measures* firm age * owned units) are positive in all four models, but only those with franchised units are significant (Models 1 and 3). This result suggests that company-franchised units rather than company-owned units drive the positive interaction effect with franchisor size.
Lastly, while we used the variables "number of years in business before starting the franchising activity" and the "location of the franchisor headquarter" as instrumental variables to control for potential endogeneity, they might also constitute meaningful control variables.
Years in business before franchising reflects the experience before franchising (El Akremi et al., 2015); headquarter location proxies some location-related variance (Combs and Ketchen, 1999b). Therefore, we estimated a separate model without controlling for endogeneity but with "years before franchising" and "headquarter location" as two additional control variables. The results remained stable. Relatedly, to assess the degree to which the set of control variables we formed affect our results, we estimated different regression models where we first used no control variables at all and then when we used our control variables in different combinations (Hayes, 2018). In all of these analyses, the results remained very consistent.

DISCUSSION
Drawing from resource-based theory and analyzing a longitudinal dataset of family and nonfamily franchisors, we find that the number of contracts canceled is significantly lower for family firm franchisors; we also find that there is no significant difference regarding the number of actions brought to court. As such, we discover partial support for our prediction that family firm franchisors build stronger relationships with franchisees than non-family firm franchisors.
Moreover, we reveal that family firm franchisors engage in more training activities with franchisees than non-family firm franchisors do. Also, our analyses confirm that non-family firm franchisors outperform family firm franchisors and that this pattern reverses when family firm franchisors are older and larger. Taken together, we demonstrate that family firm franchisors differ from non-family firm franchisors regarding franchising behavior and performance in very important ways. These insights offer valuable contributions to two literature streams: corporate entrepreneurship and franchising.

Contributions to the Corporate Entrepreneurship Literature
There is relatively little research framed around franchising as a phenomenon in the corporate entrepreneurship literature. This is somewhat surprising in that viewing franchising as a type of external corporate venturing (Sharma and Chrisman, 1999) intersects with corporate entrepreneurship as an entrepreneurial activity. Additionally, a number of scholars highlight the importance of gaining a greater understanding of the behaviors and resulting outcomes of entrepreneurial actions of individuals working in companies with different organizational forms such as franchising (see Ucbasaran et al., 2001). Through this study, we contribute to the corporate entrepreneurship literature by depicting franchising as a unique and prevalent form of firm-level corporate entrepreneurship. As such, this work adds to the existing debate about the legitimacy of the view that franchising is a viable type of entrepreneurship (Ketchen et al., 2011).
Moreover, our research provides several valuable contributions to the increasing number of studies dealing with corporate entrepreneurship in family firms (Kellermanns and Eddleston, 2006;McKelvie et al., 2014;Zahra et al., 2004). Scholars completing research in this domain are focusing on different aspects of corporate entrepreneurship such as entrepreneurial orientation (e.g., Chirico, Sirmon, Sciascia and Mazzola, 2011b;Zellweger and Sieger, 2012) or internal corporate venturing (Minola, Brumana, Campopiano, Garrett and Cassia, 2016). In this work, we seek to examine the position that franchising offers a promising and rich context to study differences in entrepreneurial behaviors and outcomes between family and non-family firm franchisors.
In particular, we contribute to academic dialogues occurring around three distinct but related issues. The first one addresses the question of whether family firms are more or less entrepreneurial than non-family firms (Eddleston et al., 2012;Randolph et al., 2017). Here, we theorize that there are differences in central aspects of franchising, with family firm franchisors establishing stronger relationships with their franchisees and providing more in class and onthe-job training to them. We interpret this as a signal of family firm franchisors exhibiting a stronger commitment and devotion to strengthen franchising as a type of firm-level entrepreneurial behavior compared to non-family firm franchisors. Accumulating resources effectively as a foundation for developing capabilities that franchisors and franchisees may use to create value for stakeholders is an outcome of this strong commitment.
The second debate refers to the resulting performance differences between family and non-family firms (Anderson and Reeb, 2003;Miller and LeBreton-Miller, 2006;Villalonga and Amit, 2006). In this regard, we theorize and empirically reveal an intriguing and novel pattern: while family firm franchisors tend to invest more in franchisees' relationships and training, this still leads them to underperform compared to non-family firm franchisors. This logic appears to indicate that their franchise-related investments are less likely to pay off in performancerelated terms. This disadvantage, however, reverses as family firm franchisors age and grow in size. As such, our theory offers novel theoretical arguments suggesting that efficient structuring, bundling, and leveraging of resources and the organizational learning that results from these processes take time and require an appropriate firm size and age in specific contexts. In particular, in line with Poppo and Zenger's (2002) work concerned with formal contracts and relational governance functions, our theoretical reasoning implies that the members of an older and larger family firm franchisor are able to efficiently couple their high levels of relational governance with increasingly customized franchising contracts that maximize performance outcomes. As a whole, our work signals that being a family firm can pay off over time in entrepreneurial and performance-related aspects. This finding challenges the conjecture that the commitment to entrepreneurial actions in family firms erodes over time and as the business grows in size (e.g., Chirico and Salvato, 2016;Naldi et al., 2007;Schulze, Lubatkin and Dino, 2003).
Relatedly, given arguments that conceptualizing corporate entrepreneurship as a specific type of strategy is appropriate (Ireland, Covin and Kuratko, 2009) To summarize, the overall picture emerging from our study is that family firms, at least family firm franchisors, appear to have a stronger commitment to developing and implementing a corporate entrepreneurship strategy -in the form of franchising -compared to non-family firms. Still, family firm franchisors underperform their non-family counterparts until they become older and larger. With these insights, we enhance our understanding of the behavioral and performance differences in franchising as a type of entrepreneurial behavior between two different organizational forms and reveal the boundary conditions under which family firm franchisors over-or underperform relative to non-family franchisors. The pattern we reveal is novel to the corporate entrepreneurship literature and calls for validation with respect to other factors (e.g., entrepreneurial orientation), practices (e.g., internal corporate venturing), and organizational forms (e.g., non-profit organizations).

Contributions to the Franchising Literature
Even though family firms play a key role in franchising (Chirico et al., 2011a;Welsh and Hoy, 2017;Welsh and Raven, 2011), research completed to examine the franchising phenomenon in this organizational form is surprisingly scarce. To contribute to the literature, we developed what we believe are novel theoretical arguments that draw a more nuanced picture of franchising.
In relation to franchising performance, to the best of our knowledge, our work is the first to offer fine-grained insights and empirical evidence that the franchisor's form of governance (family firm versus non-family firm) is a central determinant of performance. Second, and consistent with Combs et al. (2004: 916) who assert that "any relationship between franchising and performance is at best contingent on other factors," a key insight flowing from our theory and related results is that the difference between family and non-family franchisors depends on the franchisor's age and size. This insight speaks directly to previous franchising scholarship reporting mixed theoretical reasoning and results regarding the effects of age and size on franchisor outcomes (e.g., Bordonaba-Juste et al., 2011;Combs and Ketchen, 2003;Kosová and Lafontaine, 2010). Our corresponding additional analysis, where we considered the number of company-franchised units and company-owned units separately, also yields important insights. In this analysis, only the three-way interaction term with the company-franchised units is significant (see Models 1 and 3 in Table IV). This finding implies that the family firm franchisor's age is crucial for firm performance; nonetheless, in relation to the family firm franchisor's total size, the number of company-franchised units is what actually matters.
Overall, the insights flowing from our work inform our understanding of franchising and the research completed to study the phenomenon. Specifically, our results demonstrate that ignoring the family business aspect in franchising relationships and outcomes may prevent researchers from identifying important patterns.
Second, in relation to potential actions and behaviors occurring within a franchising relationship, a non-finding that deserves attention as a means of expanding our understanding of the franchisor-franchisee relationship in multiple contexts is that the number of actions brought to court does not differ between family and non-family firm franchisors. A possible explanation of this finding is that family firm franchisors, in line with our initial theoretical reasoning, are likely to establish stronger relationships with their franchisees. This, in turn, should make court action less likely per se. On the other hand, problems in the family franchisor-franchisee relationship might become more emotional than with non-family firm franchisors. This is because it is not only a business but also a more emotional relationship in that the business involves family dynamics. As a result, the parties involved might choose to rely on court actions not only for business reasons but for emotional ones as well. Collectively, emotional dynamics may thus offset the lower basic likelihood of taking court actions.
Lastly, we respond to the call for scholars to use the resource-based view of the firm as a theoretical lens to examine franchising phenomena, particularly behavioral and performancerelated outcomes (e.g., Castrogiovanni et al., 2006a;Combs et al., 2011b;Combs et al., 2004); at the same time, we expand the stream of research on resource management and orchestration (Chirico et al., 2011b;Sirmon et al., 2007;Sirmon, Hitt, Ireland and Gilbert, 2011) by investigating resource management activities that occur between firms (i.e., franchisors and franchisees). For our purpose of investigating corresponding differences between family and non-family firm franchisors, drawing from resource-based theory and the related resource management implications across organizations facilitated our efforts to develop compelling theoretical arguments.

Managerial and Social Implications
Electronic copy available at: https://ssrn.com/abstract=3532121 As a managerial implication, our study highlights that business families in the franchising sector need to be patient while pursuing desired, performance-related outcomes. Specifically, involving the family in the franchising business improves social relationships and knowledge sharing, yet it pays off financially if the business family as well as other stakeholders are truly long-term oriented in terms of waiting for the family firm franchisor to age and grow. In that sense, being a family franchisor involves hard work that leads to higher performance in the long run compared to non-family franchisors. Moreover, we demonstrate the importance of familyfirm franchisors recognizing that the resources and skills they provide to franchisees through strong relationships and training efforts become the foundations for capability development.
We believe that the most effective relationships between franchisors and franchisees are those oriented to finding ways to accumulate resources for conversion into value-creating capabilities. When this happens, the likelihood of earning positive returns increases for both franchising parties. Understanding these mechanisms and family dynamics has the potential to generate important implications for the society overall, for instance in terms of enhanced job security and financial and non-financial value creation within and across communities.

Limitations and Implications for Future Research
The limitations associated with our work suggest several issues to explore in future research studies. First, as the franchisors and franchisees in our sample are US-based, generalizability of our results in other regions and countries is an issue warranting consideration. This is an intriguing issue in that there is a host of franchise regulations in more than 30 countries outside the United States. Yet, while the extent and type of regulations may vary to some extent, the core aspects remain essentially identical. Because of this, we are confident that the main arguments we explicate in this study concerning family firm franchising in general and familyfirm franchisor behavior and performance in particular are valid across contexts. Supporting this conclusion are results reported by Falbe and Welsh (1998), who found that the franchise model (e.g. contract, fees etc.) did not change when US franchisors established operations in Canada and Mexico.
Second, because of data limitations, we do not consider succession, an event that can be difficult in family firms and can lead to firm failure when handled ineffectively. When succession occurs, it can affect franchise relationships and results strongly. Future studies should focus on understanding how to nurture and maintain value-creating relationships during times of succession. Additionally, scholars could examine the extent to which franchisors proactively provide succession training and contractual advantages to multi-generational governing franchisee families. Scholars could choose to examine succession from both a franchisor and franchisee perspective. Third, data limitations prevented us from examining generational differences in family firm franchisors. For example, third-generation family members now lead a number of McDonald's franchisees that started in the 1950s. Differences in leadership practices followed by a new generation of franchisees may affect actions taken in concert with the franchisor. In addition, because a relationship between two parties can be multifaceted and complex, using more fine-grained measures to capture it is desirable. Doing this is challenging though in that collecting the necessary data is difficult. The franchise industry, in general, is particularly "private" and avoids disclosure of information, including anonymous surveys. This is a result, in part, of litigation, government oversight, and negative press over the years.
Fourth, one might wonder if the approach of identifying family firms by checking whether two or more executives have the same family name is robust enough, particularly in case of very common surnames. In some countries such as China, for instance, we may expect this approach to have limitations. However, we believe that this is not a critical issue in our study in that we validated the status of the family firms in our sample with a very careful and comprehensive procedure using secondary materials and follow-up phone calls. Moreover, such an approach is used commonly (see Deephouse and Jaskiewicz, 2013;Gomez-Mejia et al., 2001;Miller et al., 2013). Nonetheless, scholars conducting studies in the future should try to gather additional data -or at least replicate our in-depth double-checking procedure. Relatedly, comparing family to non-family firms is challenging because making a "black or white" distinction is not always easy (Klein, Astrachan and Smyrnios, 2005). We believe though that our corresponding matched sample approach is appropriate and effective in that we not only employed a family firm dummy in our empirical analyses but also a continuous measure capturing the number of family executives. Of course, future studies investigating the family versus non-family firm distinction in more nuanced ways have strong potential to contribute meaningfully to the literature.
Fifth, while we base our arguments on the notion that franchisor age and size are associated positively with favorable reputation for a firm, we could not measure reputation explicitly in our dataset. Future studies could do so and investigate whether reputation does indeed increase with franchisor age and size (and to what degree this differs between family and non-family firm franchisors). In fact, there might be scenarios where reputation declines over time while the firm grows. Studying the outcomes associated with such a relationship between reputation and firm size has the potential to yield interesting results. Relatedly, it would be valuable to explore whether and how reputation changes in the short-and long-term when, for instance, a family firm franchisor becomes a non-family firm franchisor (or the reverse).
Sixth, while we use a set of control variables that we developed in light of results flowing from previous empirical franchising studies, data limitations prevented us from controlling for additional factors that might affect our findings. Examples of these variables include internal access to resources (Barthélemy, 2008), investment amount, fees, and royalties (El Akremi et al., 2015), exclusive territory and resale price maintenance clauses, and geographic dispersion (Combs and Ketchen, 1999b). Lastly, our measurement instrument for training provided by franchisors covers required training but not potential additional voluntary training elements.
Future studies could gather corresponding information and replicate our results with a combined or separate measure(s). Moreover, examining potential differences between required and voluntary training might lead to finer-grained insights.
In addition to highlighting research opportunities emerging because of our study's limitations, there are avenues for future studies flowing from our results. In an overall sense, we believe our study suggests a need for additional research concerned with franchising in the family firm context. In our view, there is significant research potential associated with efforts designed to examine how family-related aspects (e.g., motives, values, or family involvement in general) affect behaviors and performance in the franchising context. Consistent with specifications of the relationships analyzed in our work, we conclude that resource-based theory is appropriate to inform the specification of relationships to study regarding various familyrelated aspects in the franchising context.
Related to the issue of theory, we note that as detailed earlier in our paper, several reasons support our decision to draw primarily from resource-based theory to specify the relationships we examined empirically. Nonetheless, combining resource-based theory with others to specify franchising relationships to study is promising. For instance, combining resource-based view arguments with institutional theory insights (Combs et al., 2009) may hold promise to identify factors with the strongest influence on the performance of global franchising chains. In addition, it might be promising to examine the ownership redirection hypothesis (Oxenfeldt and Kelly, 1969) in family versus non-family firms, whereby a combination with agency theory could yield interesting relationships to test (Combs and Ketchen, 2003;Gillis et al., 2014).
Future studies may also rely on transaction cost theory (North, 1992) to study contractual franchising relationships and the related specific assets that are particularly suitable and important when analyzing franchising in family firms. It would also be interesting to examine the results of decisions by family franchisors to transition to non-family status through IPOs, MBOs, and takeovers. For instance, are family-firm franchisors more likely to become acquisition targets than non-family firm franchisors? Here, we would assume that preferences and resource-related dynamics change considerably. Such changes might affect the behaviors and performance of franchisors.
We also believe that a decision by corporate entrepreneurship scholars to focus more on the franchising phenomenon in future research efforts has the potential to yield insights. In fact, we believe that adopting an entrepreneurship lens to study franchising issues has the potential to increase our understanding of the drivers and outcomes of firm-level corporate entrepreneurship in family versus non-family firms. Lastly and relatedly, we believe that our theorizing and findings inform the analysis of other forms of corporate entrepreneurship. For instance, scholars could investigate how family-related and family-influenced resources affect behavioral and performance-related differences in the context of other entrepreneurial activities that fall into the category of "external corporate venturing" such as spin-offs or joint ventures (Sharma and Chrisman, 1999). Moreover, the mechanisms we identified could also be "at work" when corporate venturing is conducted internally (i.e., internal corporate venturing, see Minola et al., 2016).
To conclude, drawing from resource-based theory arguments, our predictions and results suggest that family firm franchisors behave and perform differently than non-family firm franchisors. We hope that our study will stimulate further examination of the differences between family and non-family firms as well as the heterogeneity within family firms in the franchising context. We believe that additional scholarship has the potential to inform our nascent understanding of family firm franchising activities and the outcomes resulting from them.   Electronic copy available at: https://ssrn.com/abstract=3532121