Many studies have focused on the role that interlocking directorates (ID) could play in affecting the performance of the firm. Most of the studies have focused on datasets constituted by samples of large firms, often selected because listed on international stock markets. If these studies have had the merit of analyzing quantitatively the effects of interlocking, the fact of being based on limited samples of companies did not allow to generalize the results to wider firm communities, especially small-sized ones. In this section, we summarize some of the results paying attention to the data and method used to obtain them.
Boyd 1990, based on a sample of 147 corporate firms listed on the stock exchange, conducted an analysis using the ordinary least square and the maximum likelihood estimator to find that firms with higher number of ties through having their directors on multiple boards exhibit better performances in terms of faster increase in sales and higher returns on capital. The underlying reason for the performance improvement is that more connected firms are more shielded from the negative effects of uncertainty.
Phan et al 2003 study the effects of interlocking for a sample of 191 listed companies in Singapore collecting information from several sources annual balance sheets and other reports, articles in the news and magazines, documents issued by the stock exchange and phone interviews. The results confirm the idea that interlocking is a way to gather information on the environment where the firm operate, alleviating risks connected to uncertainty (the authors find that their results are in line with Pfeffer e Salancik 1978).
Fich and Shivdasani 2004 conduct a panel data analysis on 508 firms from the Forbes 500 of 1992. The firms are selected based on being the larger in terms of sales, market capitalization and profits over the period 1989-1995. The fixed effect regressions show that firms with “buzy” directors, that is active on multiple boards, suffer worse performances in terms of market-to-book ratio and profitability.
Non and Frances 2007 exploit a panel of 101 large firms from the Netherlands over the decade 1994-2004 using information of the annual balance sheets as stored in the database REACH. The study confirms the negative effect of “business” of the manager, which are often part of cohesive groups of managers with multiple official assignments and meet regularly with each other’s. The results also support the idea that more cohesion slows down firms in taking decisions as they often seek unanimity, and that this might also negatively affect the formation of critical and independent thought across boards of directors (Janis, 1982, e Mullen et al., 1994).
Roomens, Cuyvers and Deloof 2007 study a dataset of 286 companies connected to a holding listed company and 2,136 independent companies in Belgium; with this dataset the authors can control for both types of connections within and outside the group types of connections. The results suggest that firms within the group do not suffer issues of business of their manager and that the interlocking does not negatively affect the firm performance (nor positively).
Koka and Prescott 2008 conduct an analysis on 422 firms forming 766 alliances across 48 countries but just with a net prevalence of firms in the metallurgic sector (166 firms). The focus of the study is on strategic alliances and the different role that firms can play within the ID network, that is as firms occupying prominent positions or just entrepreneurial positions. The factor affecting performance is indeed the role and not the degree of connectivity, firms in weaker positions tend to perform worse with respect to those occupying strategic posts. For exposition purposes, we summarize the studies presented above through a list of relevant research questions.
RQ 1: Is the ID endogenous? If the endogeneity is not properly treated, can it jeopardize the validity of the ID and performance analysis?
RQ 2: Is centrality in the ID network associated with positive effects on the performances?
RQ 3: Are there valid instruments to correct for ID endogeneity?
Among the few studies analyzing the impact of being part of a network for Small and Medium Firms (SMEs) is the paper by Wincent, Anokhin and Örtqvist 2010. The dataset is constructed to encompass 53 strategic networks of SMEs: such groups are formed on a voluntary basis by firms with the clear objective of enhancing profitability, among firms relatively geographically close; the network of firms is identified in Sweden through interviews and qualitative research of the authors. The key purpose of the study is to assess the effects on of the diversity in background and human capital within the board of the firm: diversity appear to have a clearer effect on incremental research, whereas the level of instruction tends to affect more radical innovation. A relatively negative effect on innovation is instead played by the frequency of meetings of board members. Mazzola, Perrone and Kamuriwo 2016 argue that the additional benefit in terms of innovation stems from the presence of prominent directors from prominent firms in the interlocking letting trickle down useful information to other firms in the network; being prominent firms in the network also affects positively new product development.
Santos, Da Silveira and Barros 2012 analyze an unbalanced panel of around 320 firms listed on the Brazilian stock exchange (BOVESPA) for three years to assess the impact of interlocking on performance. This study is one of the most rigorous with respect to the robustness of the results, obtained through the use of a wide set of regression models, ranging from random effects, fixed effects to two-stage least squares and the general method of moments. The main result is again that managers sitting on multiple boards tend to be associated with lower book values of the firms.
Larcker, So and Wang 2014 use information from the Corporate Board Member Magazine Director Database on 115,411 and construct a panel of 6,600 firms and 52,000 directors over the period 2000-2007. The sample is constituted by all the listed companies on stock exchanges such as NYSE, NASDAQ and AMEX, in addition to very large companies with sales exceeding $ 1 billion. The results tend to find a positive relationship between centrality in the network of the firm and future return on assets, especially for younger firms. The positive effect goes in favor of the hypothesis that connectedness implies enhanced access to resources such as information and capital.
Croci and Grassi (2014), based on OLS analysis implemented on data relating 282 firms, find a positive effect of betweenness on firm value and a negative effect of centrality measures such as degree and eigen centrality with respect to the value of the firm. As regards the effect on return on assets they find a negative result, which this time is found also for betweenness; the clustering coefficient is never significant.
Always referring to the case of Italian firms, Bellinzier and Grassi (2013) study the evolution of the interlocking network between 1998 and 2011 of listed companies on Milan stock exchange and find out that it has a persistent small world structure. The network is cohesive thanks to few directors (big linkers) that tend to sit in many boards. Persistence in the network is mainly due to the reduced number of families controlling the firms and to cross-shareholdings of companies. The analysis refers to publicly listed company and the companies occupying central positions in the network are those representing the pillars of Italian capitalism: Pirelli SpA, Mediobanca, RCS, Italcementi.
Zona, Luis and Withers (2018) address the analysis of performance referring to a resource-based and agency cost theory. According to the former conceptual framework interlocking would enhance performance as it would relax the internal constraints of the single firm allowing for access to a greater amount of resources, whereas according to the to the agency cost theory ID would exacerbate the problems relating to managerial opportunism with detrimental effects on performance. Based on GMM estimates on data relating 145 firms over the period 2001-2006, the authors conclude that the outcome in terms of performance depends upon the dyadic relationship between the focal and the interlocked firms: putting together the perspectives of agency and resource dependence shows that executives may use interfirm tactics for different purposes and that this might depend on the characteristics of the firm and of the interfirm relations. More in particular successful interlocking is those between firms with limited resources that are coopted in the network of richer and more powerful firms. As regards the implied results connected to the agency cost-based theory, the interlocking might be more problematic when the firm is more connected with fragmented ownership, that is with manager that are more independent from providers and shareholders.
Brauno, Briones and Islasa 2019 analyse the interlocking directorate of 252 firms, in practice nearly all the firms listed on the Chilean stock market, 1198 directors sitting on 1873 places across the boards of these firms. The key result of the paper is that firms interlocked with banks have superior market evaluations and tend to last longer on the market
Summarizing, several studies have been conducted worldwide to assess the effect of interlocking directorates on the performance of the firm. The most of articles have focused on large companies listed on public stock exchanges. The studies in general tend to suggest that the degree has a negative impact on the profitability and value of the firm, and this could result for instance from the business of the managers, which might lose focus from fostering the fundamentals of the firm. However, results might change when analyzing the relationship from different angles (e.g., agent based. Vs agency cost theory) or referring to specific layers of the network (e.g., prominent vs entrepreneurial firms). The studies have usually limited samples cutting out smaller firms. For the sake of clarity, we summarize the studies discussed above through two relevant questions related to the size of the firms.
RQ 4: Is the magnitude of the impact of ID on performances larger for smaller firms?
RQ 5: When ID endogeneity is not treated properly in the econometric analysis, is the bias in the ID and performance relationship larger for smaller firms?