Measuring social influence for firm-level financial performance

Abstract

Social influence is a universal concept that measures the interactions and links between entities. Existing social influence research primarily focuses on friendship networks among people. We propose a general approach to measure social influence in a type of objective entity, specifically in company networks. To construct the company influence network, we mine network links (company relationships) from the news and obtain node attributes (company influence) from search indices. Using the company influence network, two social influence measures are explored: the number of peers and weighted peer effects. To address time-series data, all variables are integrated into a vector autoregression model to forecast the company’s financial performance in terms of stock return and risk. The results of our simulation and robust testing suggest that our social influence measures have the power to predict firm financial performance.