According to the six degrees of separation theory1 , everyone on this planet is connected to each other by a maximum of six steps i.e. the number of connections needed to link to a targeted person. Hence, your friend’s friend could potentially introduce you to Roger Federer, or if he does not, then at least your friend’s friend’s friend’s friend’s friend will.
Networks are of high value for personal and business reasons. In a globalized and digitizing world, individuals are likely to be increasingly interwoven. Recent research by Facebook2 has shown that in their network members are actually separated by a mere 3.5 steps, not six.
Though meeting a superstar like Roger Federer would certainly be thrilling, in a business context it would be of greater value to be introduced to a new potential employer or employee, client, business partner, etc.
In this viewpoint, we explore the power of networking among those who serve on the Boards of Directors. We use the word “boards” here to refer such boards and “directors” to refer to board members. Unlike executives such as a CEO or CFO, or any other employee who works for a single company, directors are not constrained to having only one mandate and thus may have an interlocking relationship with multiple companies
In the context of this viewpoint we define director interlocks as a connection between at least two companies that is created when a director of one firm sits on the board of another. This link is at two levels: the firms are linked by sharing directors, and the directors themselves are linked by the direct personal interaction they have among each other.
This phenomenon means that the top decision-makers and their companies are more strongly connected than other players in the marketplace. To better understand the characteristics and functionality of a network resulting from board connections, we look at the Swiss All Share Index universe, which includes all SIX Swiss Exchange-listed shares of companies domiciled in Switzerland or in the Principality of Liechtenstein.
Each company’s board is viewed as a node in a company-to-company network. Two companies are connected if they share at least one common board member. Theoretically, there are 27’261 possible director interlocks within the Swiss All Share Index currently comprising 234 companies. Yet our analysis reveals 241 such connections.
Figure 1 depicts an excerpt of the Swiss All Share Index universe representing the maximum set of companies connected by a path, i.e. a way by which any company can be connected to any other company. Our focus is this largest connected grouping, which comprises 136 companies (or 58 percent of the Swiss All Share Index universe). We refer to this grouping as the Swiss directorship network (SDN). Based on our data, it takes 4.8 steps within an SDN on average for two companies to be connected via shared directors. The longest path contains 12 steps.
Knowing who are the members of the network is itself of significance. So is knowing the steps needed to connect to other network members. But even more important is identifying who may have greater potential to have influence in the network. A closer look at the relations within the SDN allows a better understanding of the interlocking phenomenon from an influence perspective.
One way to identify “influencers” is to consider the number of direct connections of network members to others (i.e. degree centrality). Our data shows that SDN network members have 3.4 direct connections on average.
The degree centrality score of some Boards, however, materially exceeds this average value. Figure 2 depicts companies with the highest degree centrality score. These companies enjoy the highest exposure to a network and can be referred to as network “hubs”.
The degree centrality, however, provides only local information about a network member without giving consideration to the rest of the network. Ultimately, all connections a node has are of value, though not of equal importance. A node’s importance could be assessed by its eigenvector4 centrality. Overall, a high eigenvector score means that a node is connected to many nodes which themselves have a high degree of centrality scores.
As can be observed by comparing Figures 2 and 3, boards having many direct connections do not necessarily possess a high eigenvector centrality score. Moreover, a board with high eigenvector centrality is not necessarily highly interlocked: it may have few but very important links. Expressed otherwise, information about eigenvector centrality allows one to identify a) the most central actors in terms of the “global” structure of the network and b) those more “local” actors.
Another widely used global centrality measure is closeness centrality. This measure focuses on the distance of a node to all other accessible nodes. In this manner, it allows for the identification of points of rapid information diffusion. These points can be imagined as network “broadcasters”. The lower the closeness centrality (i.e. the higher the inverse closeness score), the easier it is for a company to contact all other members of the network. Figure 4 lists such identified SDN “broadcasters”.