The Herd Effect on Chinese Firms’ OFDI - A Data Mining Approach

Abstract

When Chinese firms make OFDI decisions, their investment motives and layouts are often influenced by peers or trends, resulting in investment agglomeration and a “herd effect”, which can alter the knowledge of a host country’s risk, influence the risk attitudes of the Chinese firms, and affect their foreign investment decisions. To quantify the Chinese Outward Foreign Direct Investment (OFDI) herd effect, four dynamic indexes were developed at both industry and country levels based on a combined dataset from 2004 to 2015 for 1207 Chinese OFDI events. Data mining was then used to determine the links between Chinese OFDI volume and the herd effect and to examine the heterogeneous characteristics of the host countries, industries and firms. Using a random forest method, 50 decision attributes that affected the firms’ investment volume were identified, which were then incorporated in an optimized BP neural network to generate a Chinese OFDI decision-making model. It was found that: (1) there was an obvious herd effect in Chinese OFDI associated with host country and industry selection; (2) when a firm invested, it tended to choose a host country that has a smaller political risk and higher degree of labor freedom and globalization; (3) large firms with low efficiency tended to make larger OFDI decisions.