An Application Of An Artificial Neural Network System To Predict Takeover Targets

As a significant way to reallocate the social resources reasonably, Merger andacquisition(M&A) has always been playing an active role in both domestic and overseacapital markets. M&A has such a strategic significance that it has an irreplaceable effect onthe development of a firm. While a successful M&A deal brings rich profits to the firm, afailing one may drive the firm to the last ditch. The selecting of the target firms is especiallyan important step during the course of M&A decisions. Under different market environmentsand circumstances, the M&A motive and behavior differ from one firm to another. Among agreat number of M&A deals, how to identify target firms is quite important to either listedcompanies, investors, intermediaries or regulatory authorities, and so on. Since theimplementing reform on non-tradable shares fundamentally changes the market backgroundfor listed companies, a new M&A trend is coming. Therefore, in current new environmentand policies, all kinds of interest groups pay more attention to how to predict target firmsduring the course of M&A.This study applied variables selected based on acquisition theories and attempted toidentify potential takeover targets using all available firms from2009to2010. Also, weexamined whether a neural network could outperform Logit analysis.We found that acquired firms were usually small, experienced lower growth rate, andwere relatively concentrated in certain industries. We proved that takeover targets could bepredicted using neural network model and the results are very promising. The prediction rateis reach to78percent, performing better than logit analysis. And neural network model is lesslikely to identify non-target to target firms.