Predicting Takeover Targets in the US Technology Industry

Abstract

This thesis explores which factors affect takeover prediction in the US technology industry and whether abnormal returns are achievable with an investment portfolio based on takeover probabilities. With a sample consisting of 581 target- and 2130 non-target observations from the period 1993-2014, the takeover prediction probabilities are calculated through a logistic regression model. Incorporating the fifth and sixth merger waves in a model focusing solely on the US technology industry is new to this field of research. The results from the logistic regression indicate that (increases in) Revenue Growth along with the Current Ratio and Debt/Assets have a significantly negative impact on takeover probability, while (increases in) the Natural Logarithm of Revenue, Dividend Yield, Fed Rate and Industry Disturbances have a significantly positive impact on takeover probability. The estimates are applied on a hold-out sample consisting of 145 target- and 675 non-target observations over the period 2015-2018 to form two investment portfolios. The portfolio formed by the minimum misclassification-strategy (Palepu, 1986) achieves 2.06% abnormal return over the period, predicting 27.54% of the targets and 84.31% of the non-targets correctly. The portfolio formed according to the maximum target-strategy (Powell, 2001) achieves –5.32% abnormal return over the period, predicting 83.33% of the targets and 83.79% of the non-targets correctly. Thus, the results suggest that one can predict takeover targets quite accurately, though there are limitations to the extent to which one can achieve abnormal returns from it. This provides an exciting basis for future extensions and utilization of the industry-specific takeover prediction model. Key words: Takeover prediction, logistic regression, abnormal return, investing strategy, technology, market efficiency JEL classification: O51, L63, L65, C53, G11, G14, G34