Paper: arXiv 2404.07179
Authors: Giambattista Albora, Matteo Straccamore, Andrea Zaccaria
Abstract
Defining and finalizing Mergers and Acquisitions (M&A) requires complex human skills, which makes it very hard to automatically find the best partner or predict which firms will make a deal. In this work, we propose the MASS algorithm, a specifically designed measure of similarity between companies and we apply it to patenting activity data to forecast M&A deals. MASS is based on an extreme simplification of tree-based machine learning algorithms and naturally incorporates intuitive criteria for deals; as such, it is fully interpretable and explainable. By applying MASS to the Zephyr and Crunchbase datasets, we show that it outperforms LightGCN, a “black box” graph convolutional network algorithm. When similar companies have disjoint patenting activities, on the contrary, LightGCN turns out to be the most effective algorithm. This study provides a simple and powerful tool to model and predict M&A deals, offering valuable insights to managers and practitioners for informed decision-making.
Complexity vs Empirical Score
- Math Complexity: 6.5/10
- Empirical Rigor: 8.5/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper presents a novel, interpretable similarity metric (MASS) with moderate mathematical formulation and modifies existing economic complexity/network metrics, indicating advanced math. Empirical rigor is high due to the use of real-world datasets (Zephyr, Crunchbase), out-of-sample forecasting, and comparison against a complex baseline (LightGCN) with explicit performance metrics.
Research Flowchart
flowchart TD
A["Research Goal: Forecast M&A deals<br>using patent data"] --> B["Methodology: MASS Algorithm<br>Simplicity & Interpretability"]
B --> C["Data Inputs:<br>Patent Activity, Zephyr & Crunchbase"]
C --> D["Computational Process:<br>Tree-based Similarity Measure"]
D --> E{"Evaluate Performance<br>vs LightGCN"}
E -->|"Mass Similarity"| F["Outcome: MASS Outperforms<br>LightGCN"]
E -->|"Disjoint Patent Activity"| G["Outcome: LightGCN More Effective"]
F --> H["Key Insight: Simple, interpretable<br>tool for M&A prediction"]
G --> H
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes the MASS algorithm, a similarity measure between companies applied to patent data to forecast M&A deals.
Source sentence
We propose the MASS algorithm, a specifically designed measure of similarity between companies and we apply it to patenting activity data to forecast M&A deals.
Assumes 1
- Similarity measures usually assume symmetry, but this does not hold in reality; MASS assumes large firms acquire small startups more likely. model
Source sentence
similarity measures usually assume symmetry, meaning that the probability of firm 𝑓1 acquiring firm 𝑓2 is the same as that of firm 𝑓2 acquiring firm 𝑓1. However, this assumption does not hold in the real world, and MASS considers that it is more likely that a large, established firm acquires a small startup than the ot
Data
- Zephyr alternative commercial
Source sentence
Zephyr https://www.bvdinfo.com/it-it/le-nostre-soluzioni/dati/greenfield-investment-and-ma/zephyr and Crunchbase database (www.crunchbase.com) are used to find information about M&A deals.
- Crunchbase alternative commercial
Source sentence
Zephyr ... and Crunchbase database (www.crunchbase.com) are used to find information about M&A deals.
- PATSTAT alternative public
Source sentence
The data on technologies and patents utilized in this research is sourced from the Worldwide Patent Statistical Database (PATSTAT, https://www.epo.org/searching-for-patents/business/patstat.html), maintained by the European Patent Office (EPO).
- AMADEUS fundamentals commercial
Source sentence
Data on companies was sourced from the AMADEUS database (https://login.bvdinfo.com/R0/amadeusneo), which archives information on over 20 million companies, predominantly located in Europe.
Costs
Not applicable
Finds 2
- LightGCN performs better than MASS on the subset of 123 M&As with zero co-occurrences. performance
Source sentence
In this case, LightGCN is able to capture the hidden, higher-order similarities between companies.
- The MASS method represents state-of-the-art performance across pair, target, and acquirer prediction tasks. performance
Source sentence
The results from three different prediction exercises (pair prediction, target prediction, and acquirer prediction) show that our method represents the state-of-the-art in estimating the probability of a M&A occurring between two firms.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 7 related papers). Coverage is partial, so this is neither replication nor refutation.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (61)
- Applied mergers and acquisitions (2004)
- Lightgcn: Simplifying and powering graph convolution network for recommendation (2020)
- Random forests (2001)
- Xgboost: A scalable tree boosting system (2016)
- Sapling Similarity: A performing and interpretable memory-based tool for recommendation (2023)
- Collaborative filtering recommender systems (2007)
- Modern information retrieval: A brief overview (2001)
- The principle of relatedness (2018)
and 53 more
- Neural collaborative filtering (2017)
- Introduction to modern information retrieval (1983)
- The building blocks of economic complexity (2009)
- What percentage of innovations are patented? Empirical estimates for European firms (1998)
- Explaining the propensity to patent computer software (2005)
- How the taxonomy of products drives the economic development of countries (2014)
- The structure of the product space and the evolution of comparative advantage (2007)
- Relatedness in the era of machine learning (2023)
- Product Progression: a machine learning approach to forecasting industrial upgrading (2023)
- Which will be your firm’s next technology? comparison between machine learning and network-based algorithms (2022)
- Machine learning to assess relatedness: the advantage of using firm-level data (2022)
- Influencing Factors and Risk Control in Cross-Border Mergers and Acquisitions (2022)
- EFFECT OF MERGERS AND ACQUISITIONS ON EFFICIENCY OF INDIAN ACQUIRING BANKS: EVIDENCE FROM INDIA (2022)
- The effects of discriminatory protections on cross-border mergers and acquisitions (2022)
- Impact of COVID-19 on mergers, acquisitions & corporate restructurings (2021)
- Mergers and Acquisitions for Business Sustainability in Emerging Markets During a Vague Era: A Literature Analysis (2022)
- Comparative Analysis of Financial Performance Before and After Mergers and Acquisitions (2022)
- Matching Industries between Target and Acquirer in High-Tech Mergers and Acquisitions (2002) — link
- Beyond knowledge bases: Towards a better understanding of the effects of M&A on technological performance (2010)
- Meta-validation of bipartite network projections (2022)
- Patent analysis for supporting merger and acquisition (m&a) prediction: A data mining approach (2008)
- Pairwise acquisition prediction with SHAP value interpretation (2021) — link
- Urban economic fitness and complexity from patent data (2023)
- Prediction and visualization of Mergers and Acquisitions using Economic Complexity (2023)
- The OECD ORBIS Database (2010) — link
- A new metrics for countries’ fitness and products’ complexity (2012)
- Economics needs to treat the economy as a complex system (2012)
- Unfolding the innovation system for the development of countries: coevolution of Science, Technology and Production (2019)
- Novelty Score and Technological Relatedness Measurement Using Patent Information in Mergers and Acquisitions: Case Study in the Japanese Electric Motor Industry (2022)
- Understanding corporate coherence: Theory and evidence (1994)
- The theory of the growth of the firm (1959)
- The technological resilience of US cities (2015)
- Mapping knowledge space and technological relatedness in US cities (2013)
- Patent portfolio analysis of cities: Statistics and maps of technological inventiveness (2018)
- The emergence of new technology-based sectors in European regions: A proximity-based analysis of nanotechnology (2014)
- The role of knowledge variety and intensity for regional innovation (2014)
- Determinants of foreign technological activity in German regions-A count model analysis of transnational patents (1996-2009) (2013)
- Technology networks: the autocatalytic origins of innovation (2018)
- Structure of the Region-Technology Network as a Driver for Technological Innovation (2021)
- Complexity economics (2013)
- The atlas of economic complexity: Mapping paths to prosperity (2014)
- The geography of technological innovation dynamics (2023)
- Economic Complexity for Regional Industrial Strategies (2023)
- Technological opportunity and spillovers of R&D: evidence from firms’ patents, profits and market value (1986)
- Beyond knowledge bases: Towards a better understanding of the effects of M&A on technological performance (2010)
- Inductive representation learning on large graphs (2017)
- Semi-supervised classification with graph convolutional networks (2016) — link
- An empirical comparison of supervised learning algorithms (2006)
- Supervised learning (2008)
- Measures of the amount of ecologic association between species (1945)
- Tackling class imbalance with ranking (2016)
- Area under the precision-recall curve: point estimates and confidence intervals (2013)
- Area under precision-recall curves for weighted and unweighted data (2014)
0 of 61 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.