Identifying Politically Connected Firms: A Machine Learning Approach*
| Published date | 01 February 2024 |
| Author | Vitezslav Titl,Deni Mazrekaj,Fritz Schiltz |
| Date | 01 February 2024 |
| DOI | http://doi.org/10.1111/obes.12586 |
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 86, 1 (2024) 0305-9049
doi: 10.1111/obes.12586
Identifying Politically Connected Firms: A Machine
Learning Approach*
VITEZSLAV TITL,†,‡,§ DENI MAZREKAJ§,¶,# and FRITZ SCHILTZ§
†Utrecht University School of Economics, Utrecht University, Kriekenpitplein 21-22 Utrecht,
3584 EC, The Netherlands
(e-mail: v.titl@uu.nl)
‡Department of Economics, Faculty of Law, Charles University, Prague, Czechia
§Leuven Economics of Education Research (LEER), KU Leuven, Naamsestraat 69 Leuven, 3000
(e-mail: d.mazrekaj@uu.nl; fritz.schiltz@kuleuven.be)
¶Department of Sociology, Utrecht University, Padualaan 14 Utrecht, 3584 CH, The Netherlands
#Nuffield College, University of Oxford, New Road OX1 1NF, Oxford, UK
Abstract
This article introduces machine learning techniques to identify politically connected
firms. By assembling information from publicly available sources and the Orbis company
database, we constructed a novel firm population dataset from Czechia in which various
forms of political connections can be determined. The data about firms’ connections are
unique and comprehensive. They include political donations by the firm, having members
of managerial boards who donated to a political party, and having members of boards
who ran for political office. The results indicate that over 85% of firms with political
connections can be accurately identified by the proposed algorithms. The model obtains
this high accuracy by using only firm-level financial and industry indicators that are widely
available in most countries. These findings suggest that machine learning algorithms could
be used by public institutions to improve the identification of politically connected firms
with potentially large conflicts of interest.
I. Introduction
In the heart of the second wave of the COVID-19 pandemic, on 26 November 2020,
a controversial investigation was brought to light in a report published by the British
JEL Classification numbers: D72, D73, H83.
*The firm accounting data for this study are protected by a confidentiality agreement and we are precluded from
sharing the data with others. Interested readers can consult the corresponding author for information on how
to obtain access to the data. The code for all figures and tables is available at https://doi.org/10.5281/zenodo.
10113144. We would like to thank Climent Quintana-Domeque for his guidance and valuable suggestions,
Benny Geys, Kristof De Witte, Giovanna D’Inverno, Mark Verhagen, Lamar Pierce, and Aniek Sies for their
useful comments and suggestions and also Alice Navratilova for excellent research assistance. Deni Mazrekaj
acknowledges funding by the Research Foundation Flanders (FWO) (grant number 1257721N) and by the
European Research Council (grant number 681546). Vitezslav Titl acknowledges support from the Horizon
Europe project ‘DemoTrans’ (grant 101059288). The authors declare that they have no relevant or material
financial interests that relate to the research described in this paper.
137
©2023 The Authors. Oxford Bulletin of Economics and Statistics published by Oxford University and John Wiley & Sons Ltd.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
138 Bulletin
National Audit Office (2020). The spending watchdog found that more than half of
the public pandemic contracts (£10.5 billion) related to personal protective equipment
such as masks and protective gloves for healthcare workers, were awarded without a
competitive tender. Nearly a third of these suppliers had links to politicians or senior
officials and were referred to a ‘high priority’ channel, which was 10 times more likely
to succeed in obtaining a contract than the regular competitive channel (Conn and
Evans, 2020). Many of these suppliers had little or no experience in supplying personal
protective equipment. For instance, a contract of £108 million was awarded to a chocolate
wholesaler (Archer, 2020). In some cases, the paperwork stating why suppliers had been
selected was missing and contracts were made only after the companies had already
started the work (Pegg, Lawrence, and Conn, 2020).
Scandals involving links between politicians and private-sector firms (political
connections) are by no means isolated incidents and can be found in virtually all countries.
For instance, following a leak from the Panamanian law firm, Mossack Fonseca, the
‘Panama Papers’ revealed that the firm created thousands of shell companies for hundreds
of politicians and public officials throughout the world (Harding, 2016). Evidently, not all
entities involved in such political connections scandals are necessarily wrongdoers, but
these examples highlight the need for transparency regarding political connections. This
is especially the case given that the number of people and firms is persistently increasing,
whereas budgets for audits are either remaining stagnant or are dropping. The United
States Internal Revenue Service audited merely 0.45% of personal income tax returns in
2019, less than half of the audit rate in 2010 (Rubin, 2020).
In this article, we use supervised machine learning algorithms to predict political
connections by constructing a novel firm population dataset from Czechia. Recently,
machine learning algorithms have been found to improve predictions of many out-
comes, such as poverty (Blumenstock, 2016; Jean et al.,2016), teacher quality (Chalfin
et al.,2016), jail-or-release decisions (Kleinberg et al.,2018), Post-Traumatic Stress
Disorder (Abbasi, 2019) and even mortality (Puterman et al.,2020). Ranking among the
most corrupt countries in Europe according to Transparency International’s Corruption
Perception Index (Transparency International, 2019), Czechia is not a stranger to political
connections scandals. On 4 June 2019 for instance, Czechia witnessed its biggest political
protest since the fall of communism after the European Commission confirmed that Czech
Prime Minister Andrej Babiˇ
s had significant conflicts of interest related to his private
businesses. Specifically, his businesses received almost 20 million euros of EU agricul-
tural subsidies while being Prime Minister (de Goeij and Santora, 2019). A unique feature
of Czechia is that information on political connections is publicly available, although
scattered. Many other countries such as France, Portugal, Canada, and the USA have intro-
duced a ban on corporate donations to political parties, and information on firms’ ownership
structure and management is not available.1In Czechia, however, political donations are
allowed, and firms’ ownership structure and management can be retrieved. By employing
1Although banning corporate donations may appear as an effective policy to curb political connection at first sight,
firms can still obtain connections by having their top officers (CEO, president, chairperson) affiliated with politicians
or by politicians having equity in the firm (Faccio, 2006). These political connections are often even more difficult
to track than corporate donations, leading to even less transparency than before the ban.
©2023 The Authors. Oxford Bulletin of Economics and Statistics published by Oxford University and John Wiley & Sons Ltd.
To continue reading
Request your trial