SOCIAL CAPITAL AND NETWORK ENTREPRENEURS
Jason
C. Senjem, Syracuse University
Kira Reed, Syracuse University
Drawing on the social network literature, this paper examines how different sources of social capital affect venture performance of Hollywood film entities. We test closure and brokerage hypotheses by measuring density and constraint of the director’s network (involving eleven production roles) linking relationships over the past three director’s films to the performance of 51 film ventures produced in 2000.
Today Hollywood is the setting of one of the world’s most entrepreneurially-oriented production networks. It is made up of free agents and small and medium-sized firms led by independent directors who come together to make films project by project (Arthur & Rousseau, 1996; Jones & DePhillippi, 1996). Film projects are staffed with talent and specialists that suit the job, demanding a great amount of flexibility and coordination. Those entrepreneurs who can bring together the best talent for the expertise that is needed have the competitive edge. This paper examines the nature of the relationships among a group of entrepreneurs (i.e., directors) in forming new ventures (film entities) in the unique setting of the Hollywood network economy. Specifically we investigate how different sources of social capital affect venture performance.
It is clear that in the film venture business, being “connected” is important. However, it is the competencies that the venture entrepreneurs develop in bringing together the appropriate individuals for the project-based venture that lead to competitive advantage (Miller & Shamsie, 1996). These competencies build reputations that signal who to bring onboard for the film project. Thus, a network is formed in which the relationships that make up the network create a source of social capital for the ventures (Nahapiet & Ghoshal, 1998). To Walker, Kogut, and Shan, “the notion of social capital implies a strategy of maintaining a structure of existing relationships” (1997: 109). Leana and Van Buren further define social capital as “a resource reflecting members’ levels of collective goal orientation and shared trust, which create value by facilitating successful collective action” (1999: 539). Consequently, social capital, as part of the initial resources available to the venture entrepreneur, may have a significant impact on venture performance (Brush & Chaganti, 1999; Cooper, Gimeno-Gascon, & Woo, 1994; Eisenhardt & Schoonhoven, 1990).
Two competing theories predict different venture outcomes depending on the source of social capital in a network. On the one hand, in the film venture setting where networks are important, entrepreneurs should seek to establish as many ties as possible to take advantage of the diversity of information and to develop opportunities for future projects (Burt, 1992, 1997; Granovetter, 1973). On the other hand, these film venture entrepreneurs should seek to limit their network to select individuals, or strong ties that will reduce the risk associated with the film venture and carry them from project to project (Coleman, 1988, 1990).
In brief, these studies posit that informationally rich social networks that contain many relationships, consisting of high-status (competent and credible) participants who come from a diverse set of disciplines, can reduce the amount of time and investment required to gather information (Burt, 1992). Accordingly, such networks can serve as valuable conduits for knowledge diffusion and transfer (Coleman, 1988); they can also facilitate knowledge combinations, which can support “knowledge creating organizations” (Nonaka & Takeuchi, 1995:67) and lead to the development of intellectual capital (Nahapiet & Ghoshal, 1998).
By testing these two competing hypotheses in the social network literature we can identify the constraints and limits of the theories supporting them. In doing this, we hope to provide insight into future research exploring social capital and networks in entrepreneurial settings.
SOCIAL CAPITAL AND VENTURE PERFORMANCE
Social capital theory and structural hole theory predict two competing hypotheses regarding the formation of Hollywood film projects (i.e., organizations) and the relationships among the competing and collaborating entrepreneurs in a network economy. First, brokerage mechanisms predict that the more (diverse) weak ties individuals have the greater the venture performance. Second, closure mechanisms predict that the stronger or more dense the relationship ties the greater the venture performance. We will explore these two ideas in relation to film venture performance below.
Social capital has been defined in various ways. Broadly defined, social capital is an asset that resides in social relations and networks (Burt, 1992; Leana & Van Buren, 1999; Tsai & Ghoshal, 1998). Social capital in Hollywood film ventures comes in the form of both formal networks (film studios) and informal networks (actors, producers, cinematographers, etc.) that influence the success of the film project. Researchers refer to social capital as either relationships existing between employees and external actors (Edvinsson & Malone, 1997; Pennings, Lee & Witteloostuijn, 1998; Stewart, 1997), or relationships existing among individuals that make up the organization (Leana & Van Buren, 1999; Nahapiet & Ghoshal, 1998; Tsai & Ghoshal, 1998). These relationships are likely to differ in the value they provide to the organization. Thus, in film ventures social capital comes from networks that include relationships with organizational entities, venture financiers, and artistic and production workers.
The film venture itself comes together usually as a collaboration among director, producer, and script writer. The venture is financed up front from a film studio or other financial backers similar to venture capitalists. The director then identifies, recruits, and manages the talent who are involved in pre-production, post-production, and during production as well as controls the artistic production of the venture.
Interestingly, according to DeFillippi and Arthur a film-making venture is different from what we typically think of business enterprises in several ways. For instance, the film venture is temporary and highly variable. Indeed, the venture dissipates before even generating revenues and “no capital investments convert to fixed assets, no revenues are retained, no structure or positions are permanent, and no returns to learning accrue for future projects” (1998: 134). This variability and uncertainty of film projects enhances the necessity of using social capital to increase the performance of the venture.
Just how social capital influences performance is the subject of this paper. Two ideas offer opposing views regarding the source of social capital and its relationship with venture performance. We summarize our hypotheses below.
Network Closure Hypotheses
The network closure argument maintains that obligations, information channels, and social norms reduce risks of uncertainty thereby creating social capital (Coleman, 1988,1990). Social capital from a network closure argument is “the sum of resources, actual or virtual, that accrue to an individual or a group by virtue of possessing a durable network of more or less institutionalized relationships of mutual acquaintance and recognition” (Bourdieu & Wacquant, 1992: 119). Network closure creates an efficiency and trust of information due to the limited number of people involved in the communication (Baker, 1984) and the strength of the relationship (Podolny & Baron, 1997). In addition, network closure facilitates social norms by enforcing sanctions because “mutual acquaintances observing two people: (a) make behavior between two people public, which (b) increases the salience of reputation for entry to future relations with the mutual acquaintances, (c) making the two people more careful about the cooperative image they display, which (d) increases the confidence with which each can trust the other to cooperate” (Burt, 2000: 352).
There are a number of ways of determining network closure. First, there is network density, which refers to the strength of relationship between two individuals. In dense networks where closer relationships allow more communication, social sanctions can be more easily levied on individuals whose performance levels are below shared behavioral norms. Thus, individuals in these dense networks would be motivated to perform at acceptable standards. Thus, we predict:
Hypothesis 1a: Network density will be positively related to venture performance.
Second, there is network constraint, which is the concentration of redundant relationships within a network. The more constrained the network, the more closure the network has, and thus the greater the value of social capital is to the venture’s performance. Thus, we predict:
Hypothesis 2a: Network constraint will be positively related to venture performance.
Network Brokerage Hypotheses
Social capital from a structural holes perspective relies on the concept of weak ties put forth by Granovetter (1973). Burt (1992) summarizes the argument by emphasizing that the value of social capital lies in the information diffusion that occurs when one dense network is brokered with another. The diversity of these “ties” creates more opportunities for obtaining information, which he argues, is worth more than the value of shared norms among strong relationships.
Since network density refers to the strength of relationship between two individuals, the lower the density the more nonredundant ties there are. This nonredundancy may improve the number of opportunities to have control over and have access to diverse information that can lead to better project outcomes. Thus, we predict the following:
Hypothesis 1b: Network density will be negatively related to venture performance.
Furthermore, since network constraint is the concentration of redundant relationships within a network, the lower the density the more nonredundant ties there are. Similar to network density, this nonredundancy may improve the number of opportunities that are controlled and are accessed, thereby creating more diverse information leading to better project outcomes. Therefore, we predict:
Hypothesis 2b: Network constraint will be negatively related to venture performance.
Sample and Data Collection
The empirical data of this study is composed of relationship ties of eleven film production roles to the directors of 51 films produced in 2000 (see Exhibit 1 for a listing of the 51 film titles and their corresponding directors’ names). Data is collected on the 2000 film and compared across the three previous films of the target directors (those in the 51 films directed in 2000), thus providing a focused and defined look at the network of these film ventures. The director is considered the key entrepreneur and focus of this study because the director is the principal creative artist and typically has control over all aspects of the film.
We use data from the Internet Movie Database (IMDb), which provides data on social network and social capital variables (e.g., who works with whom). The IMDb is a database of over 260,000 film and television productions made since 1892. Created in 1990, IMDb is now an Amazon.com company. The initial sample selected for this study included 95 independent and major studio feature length films produced in the U.S. and released in 2000. However, the final sample for which data was available for all control, independent, and dependent variables for the last 3 films directed by each director, totaled 51 films and directors.
Data are gathered on eleven production roles that are part of a relationship network of the director and are involved in pre-production, post-production, and during production of the film. These roles include the producer, screen writer, lead actor, lead actress, cinematographer, film editor, casting director, production designer, art director, visual effects producer, and the film production studio. These roles were chosen because they represent Academy Award categories.
The sample includes 17 independent film studios and 34 “Major 7” film studios. Independent films typically do not have as much financing and do not have their own in-house soundstages that the “Major 7” do, which includes MGM/UA, 20th Century Fox, Sony Pictures, Warner Brothers, Paramount Pictures, Universal, and Disney. The average budget of the total sample films (independent and major 7) is $52.25 million with a range of $4 to $125 million. The average U.S. gross earnings of the total sample is $69.72 million with a range of $259,000 to $260 million. It should be noted that films often earn much more than this once the film is released abroad and/or released on video. However, to be consistent, we chose only the gross earnings for the U.S. market while the film was still showing in movie theaters.
Independent Variables
Network density. This is measured as the number of redundant relationships with the director over the last three director’s films. Specifically, the list of individuals serving in each of the eleven production roles for the film released in 2000, was compared to the list of individuals serving in each of the eleven production roles for the past three films in which the director directed (since 1995). For example, if the producer in the film released in 2000, was also a producer in any one of the director’s last three films, a 1 was entered. If the producer in the film released in 2000 had not had a previous relationship with the director in any of the director’s last three films, a 0 was entered. Thus, network density was calculated by summing across each of the eleven roles the number of individuals that had previously worked with the director. As evidenced in the descriptive statistic provided in Table 1, the mean number of relationships for each director was 5.78 with a range of 0 to 19. Theoretically, density could reach a maximum of 33 if the individuals of each of the eleven roles in the film released in 2000 were in each of the director’s prior 3 films.
Network constraint. This is a different measure of redundancy in the director’s network. This measure looks specifically at the number of times individuals in each of the eleven roles worked with the director prior to the film released in 2000. For example, the director may have worked with a particular producer twice before in two of his last three films, but did not work with this producer in the film released in 2000. Thus, this measure captures any and all redundancies in the director’s network in his last three films (prior to the film released in 2000). This measure represents a constraint to the director’s network because each additional time s/he works with the same person(s), it prohibits the inclusion of new people into the network. As evidenced in Table 1, the mean of the network constraint variable is 4.49 with a range of 0 to 15.
Dependent Variables
Venture performance was captured using an index of film ratings and awards. Film rating is measured as IMDb rating from at least 1000 online voters. The average IMDb rating for our sample was 6.65 with a range of 5.3 to 8.2. Awards is measured as the combination of the number of Academy Award nominations and number of Academy Award wins (e.g., nomination = 1; win = 2; total = 3). The average awards measure was 0.86 with a range of 0 to 12. We standardized the film rating and awards measures and conducted a principal components factor analysis using varimax rotation to see if these two performance measures captured the same thing. The two measures each loaded above 0.7 on one factor that had an eigenvalue of 1.72 and explained 57.4% of the variance. A scale reliability test indicated that the two standardized measures could be combined into one scale with a Cronbach alpha of 0.66. Thus, one performance dependent variable was created for our analysis.
Control Variables
Three control variables were chosen for our analysis: gross revenue earned, venture type, and director experience. Gross revenue is measured as total U.S. gross revenue (in millions) produced by the film venture. Venture type is measured as production backing from either a “Major 7” film studio (coded as 1) or from an independent film studio (coded as 0). Director experience measures the total number of films (non TV series) directed by a target director during his or her career. We anticipated that the better a film did at the box office, the higher the probability that it would be nominated for awards and that the public would have enjoyed it. We also anticipated that those productions sponsored by one of the “Major 7” film studios would receive more attention. Lastly, we anticipated that the more experience the director had, the greater his or her reputation, and the more likely the film would be recognized and appreciated. These three control factors are thought to be independent from either the density or the constraint of a director’s network.
Table 1 presents the descriptive statistics (unstandardized with the exception of the dependent variable) of the control, independent, and dependent variables. Two of the control variables are significantly correlated with each other: gross revenue and venture type (p<0.01). This is not surprising since the major 7 film studios invest heavily in the promotion of the films they release. The two independent variables were highly correlated above 0.8 (p<0.001) suggesting that there might be multicollinearity. However, as later discovered in our regression analysis, the VIF diagnostic procedure suggested that multicollinearity was not a problem and the VIF levels for the controls and the independent variables were under 4. Two of the control variables were significantly correlated with the independent variable, network density: gross revenue and director’s experience (each at p<0.05). In other words, higher levels of gross revenue and director’s experience were associated with a higher number of relationships in the director’s network. This finding is not surprising, although the causality of the relationship between these variables is unknown. The dependent variable, the standardized scale combining the IMDb rating and the total number of Academy Award nominations and wins, was only significantly correlated with the control variable gross revenue (p<0.05) and not with either of the independent variables. This preliminary test of the relationship between the dependent and independent variables suggests that none of the hypotheses would be supported. However, a hierarchical linear regression was used to test hypotheses 1 and 2.
The control variables (i.e., gross revenue, venture type, and director’s experience) were entered as step 1 (noted as Base Model in Table 2). The two independent variables, network constraint and network density were entered in step 2 (noted as Network Effects in Table 2). We expected to find support for hypotheses 1a and 2a if the coefficients for both network constraint and network density were positive and significant in step 2 of the model. We expected to find support for hypotheses 1b and 2b if the same coefficients were instead found to be negative and significant in step 2 of the model.
In the first step, the overall model approached significance with an F of 2.66 (R2 = 0.38; adj R2 = 0.15; DELTA R2 = 0.15; df =50; sig = 0.06). Gross revenue was the only control variable that significantly predicted the dependent variable (beta = 0.42; p<0.01). The two independent variables entered in step 2 explained a significant amount of incremental variance and resulted in a significant overall model with an F of 3.85 (R2 = 0.55; adj R2 = 0.30; DELTA R2 = 0.15; df =50; sig = 0.005). Both of the independent variables were significant predictors of the dependent variable: network constraint (beta = -0.72; p<0.01); network density (b = 0.52; p<0.05). The results of the second model suggest that a director’s network ties are significant predictors of a film’s performance. However, in the case of network density, the positive coefficient provides support for hypothesis 1a, while for network constraint, the negative coefficient provides support for hypothesis 2b.
The contrasting results of our analysis suggest that there is merit to both network closure and network brokerage arguments. Specifically, the size of an individual’s network (how many people s/he knows) is important, but size alone is insufficient to guarantee success. The true value of relationships lies in their duration and the ways in which they are leveraged. In other words, relationships are path dependent.
In the case of our study, the number of prior relationships directors have leveraged with producers, actors/actresses, screenwriters, cinematographers, film editors, casting directors, production designers, artistic directors, visual effects supervisors, and film studios positively predicts current film success more so than the director’s own prior experience or ability. Thus, establishing relationships and leveraging them at another time in the future is effective.
However, there is a diminishing return to the value that any particular network can provide. This is evidenced by the negative coefficient for network constraint. It appears that working with the same network again and again has a negative affect on the performance of a film. Thus, it may be possible to exhaust the value of a network because too many redundant ties limit the flow of new information, ideas, and creativity. Therefore, network brokerage and closure theories may represent ebbs and flow in network lifecycles: (1) in the beginning it is important to make as many contacts as possible; (2) later on it is important to leverage these relationships and capitalize upon the synergies created by a particular group of people working together repeatedly; and (3) the benefits provided by a particular network peaks until new individuals are added and dropped by the network.
Therefore, entrepreneurs should be aware of both brokerage and closure sources of social capital. The implications of our results are that entrepreneurs will find a limit as to how much they can leverage network closure as a source of social capital. For instance, when raising financing, generating sales, and co-opting resources, entrepreneurs should not exhaust their current network but extend their network through brokerage mechanisms.
Limitations
Some limitations should be recognized. First, the data should be interpreted with caution outside of the sample setting of the Hollywood network economy. This is a unique setting in that organizations are created and dissipated in a short period of time. Therefore networks in this setting, compared to traditional firms, may have different dynamic qualities. Second, we limited our sample to film ventures that had directors with experience in at least three other film projects and had been among the top 250 rated U.S. films in 2000 with at least 1000 votes, which may have restricted the range of film rating performance. Specifically, lower rated films tended to be directed by inexperienced directors that were not considered because they had not yet directed three films. Third, two network measures were used. Other network measures might be able to tell a different story.
Future Research and Implications
There are several avenues for future research. Curvilinear relationships were not tested in this study but could lend some useful insight into the opposing results that were found. Specifically, the relationship between network constraint and film performance may be curvilinear requiring a more sophisticated type of statistical method to test this relationship. This can be overcome in future studies with more diverse data collection and by testing the relationships using a variety of statistical methods.
Even within the Hollywood film industry, there is much to learn about the relative importance of the different types of relationships. In other words, how much of a film’s success is due to a director’s leveraging of past relationships with producers versus past relationships with one of the other ten remaining production roles? Consequently, the concept of network hierarchy should be explored in future research. Finally, testing the value of redundant relationships in more traditional entrepreneurial settings such as new technology dot.com firms, or family-owned restaurants, etc. will be important in extending these results to other settings.
This paper contributes to our knowledge in several ways. First, it examines the nature of relationships among a group of entrepreneurs in forming new ventures (film entities) in the unique setting of the Hollywood network economy. Second, by testing two competing hypotheses in the social network literature we can identify the constraints and limits of the theories supporting them. Finally, this may provide insight into future research exploring social capital and networks in entrepreneurial settings.
This paper provides a look into a unique network setting, that of the Hollywood network economy. The creation and production of these project-based film ventures illustrates the importance of social capital. We examined two opposing views regarding sources of social capital and its impact on venture performance. Results showed evidence for both viewpoints, which leaves an interesting story to tell for this paper and for future research.
CONTACT: Jason C. Senjem, Syracuse University, School of Management, Syracuse, NY; (T) 315-443-9610; (F) 315-443-5457; jsenjem@syr.edu
Arthur, M. B., & Rousseau, D. M. (1996). The Boundaryless Career. New York: Oxford University Press.
Baker, W. (1984). The social structure of a national securities market. American Journal of Sociology, 89 (4): 775–811.
Bourdieu, P., & Wacquant, L. J. D. (1992). An Invitation to Reflexive Sociology. Chicago: University of Chicago Press.
Brush, C., & Chaganti, R. (1999). Businesses without glamour? An analysis of resources on performance by size and age in small service and retail firm. Journal of Business Venturing, 14: 233–257.
Burt, R. (1992). Structural holes: The social structure of competition. Cambridge: Harvard University Press.
Burt, R. (1997). The contingent value of social capital. Administrative Science Quarterly, 42: 339–365.
Burt, R. (2000). The network structure of social capital. In B. M. Staw (Ed.), Research in Organizational Behavior, 22: 345–423.
Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94: S95–S120.
Coleman, J. S. (1990). Foundations of Social Theory. Cambridge, MA: Harvard University Press.
Cooper, A., Gimeno-Gascon, J., & Woo, C. (1994). Initial human and financial capital as predictors of new venture performance. Journal of Business Venturing, 9: 371–395.
DePhillippi, R. J., & Arthur, M. B. (1998). Paradox in project-based enterprise: The case of film making. California Management Review, 40(2): 125–139.
Edvinsson, L., & Malone, M. S. 1997. Intellectual capital: Realizing your company’s true value by finding its hidden brain power. New York: Harper Business (a division of Harper Collins Publishers).
Eisenhardt, K., & Schoonhoven, C. (1990). Organizational growth: Linking founding teams, strategy, environment, and growth among U.S. semi-conductor ventures. Administrative Science Quarterly, 28: 274–291.
Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78: 1360–1380.
Jones, C., & DePhillippi, R. J. (1996). Back to the future in film: Combining industry- and self-knowledge to meet career challenges of the 21st century. Academy of Management Executive, 10(4): 89–103.
Miller, D., & Shamsie, J. (1996). The resource-based view of the firm in two environments: The Hollywood film studios from 1936 to 1965. Academy of Management Journal, 39(3): 519–543.
Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and organizational advantage. Academy of Management Review, 23 (2): 242–266.
Nonaka, I, & Takeuchi, H. (1995). The Knowledge Creating Company. New York, NY: Oxford University Press.
Podolny, J., & Baron, J. (1997). Resources and relationships: Social networks and mobility in the work place. American Sociological Review, 62: 673–693.
Richter, A., & Schaffer, M. (1996). The performance of de novo private firms in Russian manufacturing. In S. Commander, Q. Fan, and M. Schaffer (Eds.), Enterprise restructuring and economic policy in Russia, 253–274. Washington, DC: World Bank.
Stewart, T. A. 1997. Brain power: Who owns it. . . . How they profit from it. Fortune, Vol. 135, 5: 104–110.
Walker, G., Kogut, B., & Shan, W. (1997). Social capital, structural holes, and the formation of an industry network. Organization Science, 8(2): 109–125.
2002 Babson College. All Rights Reserved. Last Updated March 2003.