M&A Prediction Model-Based Investment Strategies

Abstract

In this paper, we study the development of investment strategies by predicting M&A deals using a logistic model with the financial and non-financial indicators of public companies. A random sample of 1510 acquired and non-acquired companies in Germany, the United Kingdom, France, Sweden, and Russia over the period 2000-2021 was used to design an M&A logit prediction model with high predictive power. The use of interaction variables significantly improved the model’s predictive power and allowed it to obtain more than 70% of correct out-of-sample predictions. Then the model’s ability to generate abnormal returns was tested with the help of an event study using share price data over the period 2011-2021. We show that an M&A prediction model can also efficiently generate abnormal returns (up to 49% on average) for a portfolio of companies that are expected to be acquired. Moreover, we uncover evidence that reduction in false positive
and negative predictions has a positive effect on abnormal returns due to the added model flexibility resulting from interaction terms. Our positive theoretical and empirical results can help both private and institutional investors to design investment strategies. In addition, there are indirect implications that support the practical importance of an efficient M&A prediction model.