European Management Review

TABLE 1

FROM:

Comparing small world statistics over time and across countries: an introduction to the special issue comparative and transnational corporate networks

Bruce Kogut and Mariano Belinky

BACK TO ARTICLE

Table 1. Interlocking among firms, 1990 panel

Country Sample year Number of nodes Number of links % of nodes in giant component Density Clustering coefficient Average path length
Germany1993304227598.00.0490.392.6
Sweden1990229113896.50.0440.423.0
Denmark1993313743794.90.1520.732.5
Norway1990537577.40.0540.484.2
Switzerland199096392100.00.0860.412.6
The Netherlands19971152440385.20.0070.564.0

 This table presents the descriptive data for the networks of the board interlocks for six countries for the early 1990s (with the exception of the Netherlands). These data are for the one-mode projection of boards (firms rather than directors, to be clearer); the links are undirected. The nodes are the number of boards; the links are the number of interlocks. The giant component is the largest sub-graph, that is, the sub-network that has the most boards that are connected to each other either directly or by a path. The percentages are high, showing that all of these networks consist of a very large number of boards that can be reached from one another. Density is a conventional network measure of the number of realized links among nodes (boards) divided by the potential maximum of these links. The clustering coefficient is a measure of average local density. It is defined in our analysis as (three times) the number of closed triangles over the number of all nodes having two links (two-stars). Average path length is the average length of the shortest distance (geodesic) among all pairs of nodes in the giant component. Together, the clustering coefficient and path length are critical inputs into the calculation of the small world statistic defined in the text and figures below.

BACK TO ARTICLE