Merger Premium Predictions Using a Neural Network Approach

Abstract

Over 80 percent of mergers fail to achieve projected financial, strategic, and operational synergies (Marks and Mirvis 2001). It is critical for management to find accurate models to price merger premiums. Management has an interest to protect stakeholders by acquiring companies that can add value to their investments at the most favorable price. Published studies in the area of pricing mergers have not attempted to use expert systems in the decision-making process. This paper is the first of its kind that describes the development and testing of neural network models for predicting bank merger premiums accurately. A neural network prediction model provides a tool that can filter through noise and recognize patterns in complicated financial relationships. The results confirm that a neural network approach provides more explanation between the dependent and independent financial variables in the model than a traditional regression model. The higher level of accuracy provided by a neural network approach can provide practitioners with a competitive advantage in pricing merger offers.