Essays on Common Ownership and Corporate Control

Kilcioglu, Beril (2023) Essays on Common Ownership and Corporate Control. Doctoral thesis, University of London: London Business School.

Abstract

This dissertation is composed of three essays which study corporate ownership and control. The first essay revisits Azar et al. (2018a), where authors use evidence from the U.S. airline industry to substantiate the claim that common ownership -the simultaneous ownership of equity of competing companies by the same investor- leads to reduced competition. Their main results overstate common ownership since they assume pre-bankruptcy equity holdings to be continued during bankruptcy, which affects 28 out of 56 quarters of their analysis. First, I replicate their results in the extended sample (2001Q1-2017Q4) and find that increased common ownership is associated with a 6.18% increase in the average air fares when shareholders are assumed to hold their pre-petition shares and retain their pre-petition control rights. Then, I address the challenge of accounting for the multiple periods in which major airlines went through bankruptcies and investigate and show that the anti-competitive effect declines to 3.86% when index funds are forced to sell bankrupt shares, and it is neither economically large nor statistically significant when all shareholders lose their pre-petition control rights.

In the second essay, I study the simultaneous ownership of equity of competing companies by the same investor, i.e., common ownership, in 39,867 publicly listed corporations in 125 countries between 2000-2020. I construct a global data set to document the prevalence of common ownership in multiple forms, including i) common ownership by asset managers; ii) crossholdings between competing corporations; iii) business groups. Almost half of the world’s publicly listed firms have a common owner, and both the measures of common ownership and the types of common owners are heterogeneous across countries. I study the relationship between common ownership and market power. In the pooled sample, firm markups increase with the size of common ownership. This increase is even higher when common ownership across industries is accounted for. When direct shareholdings are analysed, common ownership is more prevalent in Europe and the U.S., mainly attributable to direct ownership by asset managers. Product market regulation does not seem to reduce the anti-competitive effects of common ownership. When ultimate ownership is included, we observe more concentration of ownership in Asia and outside Western Europe; but in those countries we observe less common ownership within an industry and more common ownership across industries. As a result, common ownership in those countries presents less of a threat to competition.

The third essay is joint work with Julian Franks and Elias Papaioannou that reveals the activity in the market for corporate control to be much higher than the extant literature reports. We employ big data techniques and undertake a massive data collection and cleaning task to trace the ultimate owners of world’s listed corporations in 2007, 2012 and 2019. We observe the evolution of corporate control in the balanced sample of 14,346 corporations in 95 countries. Although corporate control is persistent and stable across time, transitions between seven corporate control categories (widely held without blockholders, widely held with blockholders, government-controlled, family-controlled, bank-controlled, controlled by a non-bank financial firm, other-controlled) and changes in the ultimate owner are pervasive. We quantify “corporate control mobility” as the total percentage of firms that have gone through a change in control. There is an active market for corporate control: 21% of the global sample of listed firms transition from widely held to controlled and vice versa. The mobility is far more pronounced between control categories: 36.6% move across seven subcategories, and more than 40% change ultimate owners. Corporate control mobility is persistent throughout the sample and across geographical regions.