Many scholars have applied a network perspective when studying boards of directors (Burt, 1978; Mizruchi, 1996; Zajaz, 1988). A network is a system whose elements (or nodes) are connected through ties (Wasserman & Faust, 1994). An investigation into the boards of directors can be conducted in two ways. First, it is possible to study the interdependence among companies. For example, Levine (1979) studied how companies were connected by having the same individual on both their boards of directors. By being connected, communication, knowledge transfer, and coordination are facilitated between companies (Zajaz, 1988). It has been found that firms which are highly embedded in the network of firms are more likely to survive (Uzzi, 1996). Second, it is possible to study the interpersonal relationships among individuals that are directors on a set of boards. In this case, individuals can be linked by being directors on the same boards. Each director is assumed to possess unique knowledge, and by being connected to others, directors gain exposure to their knowledge. In this paper, we focus on the interpersonal network among directors.
Several scholars have argued for the benefits linked to certain types of networks (Burt, 1992; Freeman, 1978; Opsahl et al., 2010; Uzzi, 1996; Wasserman & Faust, 1994). While human capital is referring to individual qualities, social capital exists in relations among people (Coleman, 1988). Simmel (1950) was among the first scholars to explore which positions in a social network are advantageous. He argued that people in a position of tertius gaudens, or powerful third-party, gained some form of capital. Burt (1992) formalized this idea through the theory of structural holes. An individual can create structural holes by being connected to people who are themselves disconnected (i.e., broker between them). The theory assumes that an individual gains some form of social capital by having the ability to, for example, control the information flow between the others in a network. As Burt's (1992) brokerage measures are defined using the existence of ties among an individual's contacts, the measures are not well-suited for co-occurrence network data, such as the interpersonal network among directors (Opsahl, 2012; Wasserman & Faust, 1994). This is due to the fact that many ties exist among individuals' contacts in these networks by construction (e.g., the directors of a single board are all connected, and thereby form a fully-connected clique). However, the extent to which an individual is in 'the thick of things' and is able to broker can also be quantified by centrality measures (Freeman, 1978). One such measure is betweenness centrality, which assesses whether, and the extent to which, a node funnels the flow among others. In a social network, Freeman (1978) assumed that a piece of information flows between two people along the shortest path (i.e., the path that includes the least number of intermediate people). In turn, this assumption implies that the intermediate people are in a position of control. Formally, he defined the betweenness score of an individual as the number of shortest paths between others that passes through that person.
In a similar vein as with prominence, control over the flow of information among directors can be seen as a proxy of influence. Thus, it is interesting to study the evolution of social capital as this might disagree with the Norwegian Government's goal of equality among directors. To investigate whether the gender representation law has led to a dispersion of social capital or increased a select few directors' social capital disproportionally, we hypothesize the following:
Hypothesis 3a: The maximum level of a single individual's social capital has not increased.
Women and Social Capital
Women and men may have distinct interaction patterns, and might form and utilize their personal networks in different ways. According to Kanter (1977), male dominated power structures, especially in the corporate world, are reinforced by men choosing men of similar background and interests. Nevertheless, as the gender representation law forced a change within the boardroom, directors' networks and their level of social capital would be affected. To assess the extent to which social capital is associated with a particular sex, we put forward the following hypothesis:
Hypothesis 3b: Women and men have equally levels of social capital.
In order to study whether the gender representation law has achieved a more equal setting and even distribution of power, we investigated whether the maximum level of social capital that a single director possesses has increased substantially, and whether similarity among directors' social capital has increased during the implementation period of the gender representation law. We constructed the network each month of all the directors that were a member of at least one board on the first of that month, and linked two directors if they were part of the same board. Given that only prominent directors are able to funnel information in the network, we calculated their betweenness score (Freeman, 1978). Although both the maximum and standard deviation fluctuate each month, the values are roughly stable from September 2002 until the implementation period. Conversely, the values soar during the implementation period. As the maximum betweenness and standard deviation increase instead of decrease, hypothesis 3a is also not supported. Our findings suggest that inequality has increased rather than decreased as a result of the law.
The Figure below shows the average betweenness score for prominent women and men. These measures shed light on the effect of the law on each sex's average level of social capital. We found that in 2004 when approximately only 23% of directors were women, the average betweenness score of women was roughly twice that of men. This was followed by a sharp decline towards parity at the end of 2005. When the law was introduced in 2006, both women and men's betweenness scores rose; however, the rate of increase in women's betweenness score was much greater than men's. More specifically, women's betweenness score was approximately 70 percent larger than men's throughout 2009. As women and men do not have an equal level of social capital, we reject hypothesis 3b.The average betweenness score for prominent women and prominent men. We chose to calculate the averages on the prominent directors as the non-prominent ones do not have a betweenness score by definition.
Burt, R. S. (1992). Structural holes: the social structure of competition. Cambridge, MA: Harvard University Press.
Coleman, J. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94, S95-S120.
Freeman, L. C. (1978). Centrality in social networks: conceptual clarification. Social Networks, 1(3), 215-239.
Kanter, R. M. (1977). Men and women of the corporation. New York City, NY: Basic Books.
Levine, J. (1979). Joint-space analysis of "pick-any" data: analysis of choices from an unconstrained set of alternatives. Psychometrika, 44(1), 85-92.
Mizruchi, M. S. (1996). What do interlocks do? an analysis, critique, and assessment of research on interlocking directorates. Annual Review of Sociology, 22, 271-298.
Opsahl, T. (2012). Triadic closure in two-mode networks: Redefining the global and local clustering coefficients. Social Networks 34; arXiv:1006:0887
Opsahl, T., Agneessens, F., Skvoretz, J. (2010). Node centrality in weighted networks: Generalizing degree and shortest paths. Social Networks, 32(3), 245-251.
Simmel, G. (1950). The sociology of Georg Simmel (KH Wolf, trans.). New York City, NY: Free Press.
Uzzi, B. (1996). The sources and consequences of embeddedness for the economic performance of organizations: the network effect. American Sociological Review, 61(4), 674-698.
Wassermann, S., & Faust, K. (1994). Social network analysis: methods and application. Cambridge, MA: Cambridge University Press.
Zajaz, E. J. (1988). Interlocking directorates as an interorganizational strategy: a test of critical assumptions. Academy of Management Journal, 31(2), 428-438.
To test our hypotheses, we collected a list of all the 384 public limited companies in Norway (Allmennaksjeselskap or ASA) that were available online through the Norwegian Business Register’s website on August 5, 2009. We chose these companies as they are the ones bound by the gender representation law. Based on the list of companies, we collected all official announcements made to the register that were online. These announcements contain changes to the composition of the boards of directors since November 1, 1999. Since all companies did not change their boards immediately after the Register started publishing the announcements online, our observation period only starts in May 2002, and extends to August 2009. The choice of starting the observation period in May 2002 is based on a trade-off between the inclusion of companies and the length of the observation window. Of the 384 companies, 196 were incorporated after November 1, 1999. These companies are included in our dataset on the first of the following month of their incorporation. The additional 188 companies (incorporated before November 1999) changed their boards at various times. We chose to start the observation period in May 2002 as 90 percent of these companies had at least changed their board once by that time. Thus, information on their board compositions was published online. The remaining 10 percent (19 companies) are included in the dataset as soon as they changed their boards’ compositions. Companies that filed for bankruptcy are removed from the sample in the month following such an announcement.
From the board compositions, we extracted a list with the names of all directors. From this list, we excluded employee representatives as the legislation does not affect them in the same way. Since mistakes could have occurred while entering the data and people may change their name, this list was manually cleaned by studying the compositions over time and comparing changes. For example, Alexandra Bech Gjørv was a director of Schibsted ASA from 2001 until 2007. However, in 2001 and 2002, her name was listed as simply Alexandra Bech. Without the manual cleaning, she would have been included as two separate people in the dataset.
To determine the sex of the directors in our dataset, we collected lists of all male and all female first names belonging to more than 200 people in Norway from Statistics Norway. We cross-referenced these two lists with the first names of the directors. However, some first names were not in either of the lists, and some first names were included in both lists. In an effort to avoid having missing data, we conducted a web search to determine the gender of directors with these names.
Note: The analysis provided does only consider the data from May 2002 to August 2009 (peer-reviewed).