Modelling Market Reactions to M&A: Cash and Stock Differences Through Classical and Machine Learning Techniques

This thesis looks at how financial markets respond when a merger or acquisition is announced and tries to understand whether different modelling approaches can capture these reactions in different ways. The first part relies on the classic event study framework and on traditional econometric tools that are the approaches that remain the standard reference for measuring abnormal returns around the announcement date and for interpreting how investors absorb new information. These methods provide a structured way of studying the immediate market reaction and create a benchmark against which more flexible models can be compared. Alongside this, the thesis explores a set of machine learning techniques, which are becoming increasingly common in empirical finance and which may offer a different perspective when the underlying relationships are not strictly linear or when the effects of certain deal characteristics interact in more complex ways. A central element of the work is the distinction between cash and stock payments in M&A deals, since the literature shows that investors often react differently depending on how the transaction is financed. Building on this idea, the thesis develops an empirical setup that applies both the traditional models and the machine learning ones to the same dataset, so their behavior can be compared in a consistent and meaningful way. This comparison aims to highlight not only the strengths and limitations of each method, but also the kinds of patterns they tend to emphasise when analysing market reactions. The goal is not to replace the conventional approach with a more modern one, but to understand what each technique is able to capture, where their interpretations overlap, and where they do not, especially in relation to how markets process and evaluate M&A announcements. By bringing these perspectives together, the thesis seeks to offer a clearer view of the tools available for studying corporate events and of the insights that different modelling strategies can provide.