For more information visit Skatteforsk - Centre for Tax Research.
Abstract
A large literature considers the interplay between mergers and acquisitions
(M&A) by multinationals and innovation, typically measured as patents. One theme that emerges is that technological similarity plays a role in which M&As occur. The challenge then becomes how to measure technological similarity. Typically, this is done by using either firms’ sectors or the technological classifications of their patents. The first is a coarse measure of business activity and overlooks supply chains. The second may incorrectly match technical features with no overlapping business relation. In contrast, we offer an alternative based on a machine learning approach using patent technology descriptors. This decants over 600 technological codes into 21 technical business areas (TBAs). Similarity in this more parsimonious measure is highly related to the probability of a given M&A occurring. This provides a more intuitive approach to understanding the literature’s findings.
Keywords: Mergers and Acquisitions; Technological Similarity; Multinationals; Patents.
JEL classification: F23; G34; O30