Corporate Control around the World

• Published date: 01 June 2020
• Author: ELIAS PAPAIOANNOU,GUR AMINADAV
• DOI: http://doi.org/10.1111/jofi.12889
• Date: 01 June 2020

THE JOURNAL OF FINANCE •VOL. LXXV, NO. 3 •JUNE 2020

Corporate Control around the World

GUR AMINADAV and ELIAS PAPAIOANNOU∗

ABSTRACT

We study corporate control tracing controlling shareholders for thousands of listed

firms from 127 countries over 2004 to 2012. Government and family control is per-

vasive in civil-law countries. Blocks are commonplace, but less so in common-law

countries. These patterns apply to large, medium, and small firms. In contrast, the

development-control nexus is heterogeneous; strong for large but absent for small

firms. Control correlates strongly with shareholder protection, the stringency of

employment contracts and unions power. Conversely, the correlations with creditor

rights, legal formalism, and entry regulation appear weak. These patterns support

both legal origin and political theories of financial development.

UNDERSTANDING THE DETERMINANTS AND consequences of the various types of

corporate control is of first-order importance in corporate finance (La Porta,

Lopez-de Silanes, and Shleifer (1999), Tirole (2006)). Although most theory

distinguishes between widely held corporations with dispersed ownership and

firms in which a dominant shareholder exerts control (Shleifer and Vishny

(1997)), corporate structures are complex (Laeven and Levine (2008)). For

instance, equity blocks can be found in most widely held corporations (Edmans

and Holderness (2017)). In addition, pyramids that allow shareholders to influ-

ence decisions over their cash flow rights and cross-holdings of equity in busi-

ness groups are widespread (Dyck and Zingales (2004)). Ownership and control

are often obscured by companies that incorporate offshore centers (Zucman

(2015)). Following the influential contribution of La Porta, Lopez-de Silanes,

and Shleifer (1999), a voluminous literature studies ownership concentration

and corporate control across countries, with particular emphasis given to the

role of investor protection rights and legal origin. To date, however, the litera-

ture has yet to reach a consensus even on the basic correlations, as researchers

face a number of empirical challenges. Understanding the determinants and

∗Gur Aminadav is with London Business School. Elias Papaioannou is with London Business

School and CEPR. We would like to thank two anonymous referees, the Associate Editor, the

Editor (Amit Seru), Julian Franks, Rafael La Porta, Sebnem Kalemli-Ozcan, Richard Portes, and

Andrei Shleifer for useful comments and valuable feedback. We would also like to thank Sebastian

Hohmann, Irene D´

ıaz de Aguilar Hidalgo, Giorgio Chiovelli, and Lu´

ıs Fonseca for their help on

various parts of the paper. All errors are our sole responsibility. We have read The Journal of

Finance disclosure policy and have no conflicts of interest to disclose.

Correspondence: Elias Papaioannou, London Business School, Regent’s Park, London, NW1

4SA, UK; e-mail: eliaspapaioannou@london.edu.

DOI: 10.1111/jofi.12889

C2020 the American Finance Association

1191

1192 The Journal of Finance R

consequences of the various types of corporate control is of first-order impor-

tance in of corporate finance (La Porta, Lopez-de Silanes, and Shleifer (1999),

Tirole (2006)). Although most theory distinguishes between widely held corpo-

rations with dispersed ownership and firms in which a dominant shareholder

exerts control (Shleifer and Vishny (1997)), corporate structures are complex

(Laeven and Levine (2008)). For instance, equity blocks can be found in most

widely held corporations (Edmans and Holderness (2017)). In addition, pyra-

mids that allow shareholders to influence decisions over their cash flow rights

and cross-holdings of equity in business groups are widespread (Dyck and

Zingales (2004)). Ownership and control are often obscured by companies that

incorporate in offshore centers (Zucman (2015)). Following the influential con-

tribution of La Porta, Lopez-de Silanes, and Shleifer (1999), a voluminous liter-

ature studies ownership concentration and corporate control across countries,

with particular emphasis given to the role of investor protection rights and

legal origin. To date, however, the literature has yet to reach a consensus even

on the basic correlations, as researchers face a number of empirical challenges.

The first challenge relates to sample size and composition. In particular,

because it is hard to identify control from a myriad of complex corporate own-

ership structures, comparative studies typically work with samples covering

large firms in a few countries. La Porta et al. (1999) examine the association

between legal origin and control for the 20 largest listed firms in 27 advanced

economies. Extending the analysis, Claessens, Djankov, and Lang (2000) study

the association between legal origin/institutions and ownership for 2,980 firms

in nine East Asian countries, Faccio and Lang (2002) look across 5,232 firms in

13 Western European countries, and Lins (2003) works with 1,433 firms in 18

emerging markets. However, even when Holderness (2016a,2016b) reassesses

theLaPortaetal.(1999) results merging these data, he only works with ap-

proximately 8,000 firms in 32 countries.

Heterogeneity is the second main challenge. The size distribution of listed

firms is highly skewed. In 2012, for instance, the average market capitalization

in our sample is 10 times larger than the median. Verylarge, medium, and small

listed firms differ along numerous dimensions (see Gabaix (2009,2016)) and

control patterns are likely to vary widely (Tirole (2006), Holderness (2016b)).

Third, researchers face measurement challenges. It is difficult to identify

controlling shareholders from a complex network of equity holdings and or to

identify control in firms with multiple large shareholders. The patterns may

depend, for instance, on the cutoff that researchers use to identify control (e.g.,

Holderness (2009)). In addition, while many researchers work with databases

on cash flow, looking at voting rights is conceptually more appealing. Prior

studies also show that institutional proxies suffer from measurement error

(e.g., Glaeser et al. (2004)).

In this paper, we make progress on each of these fronts. We first provide

a comprehensive description of corporate control for a wide sample of coun-

tries and listed firms. Relying on numerous sources (e.g., regulatory filings,

company reports, government publications), we augment Bureau van Dijk’s

(BvD) ORBIS database on corporate ownership to identify ultimate controlling

1193

shareholders from the complex structures of corporate holdings. We apply both

a simple cutoff approach that identifies as controlled those firms in which a

shareholder (state, family, other) has more than 20% of the voting rights and

an alternative game-theoretic method based on the Shapley-Shubik (1954) vot-

ing power index to construct measures of corporate control for 42,720 listed

firms from 127 countries over the period 2004 to 2012.1Given the wide use

of equity blocks, we distinguish between three types of firms: widely held cor-

porations, widely held corporations with one or more equity block(s) (defined

as voting rights in excess of 5%), and controlled firms with a dominant share-

holder. We split controlled firms into state-controlled, family-controlled, and

controlled by other listed or private firms. We provide an anatomy of corporate

control using these newly compiled data. The descriptive analysis reveals large

differences in corporate control around the world, a result that is consistent

with earlier studies that work with smaller and less representative samples.

We further show that corporate control patterns are persistent, as the 2007 to

2010 financial crisis did not affect them much.

Second, we reexamine the “reduced-form” correlation between corporate con-

trol (and ownership concentration) and legal origin. The large sample is useful,

as most previous studies consider smaller firm samples with limited country

coverage. The large sample is also helpful in examining heterogeneity with re-

spect to firm size and age, characteristics that may affect control and in turn be

affected by the institutional environment (Foley and Greenwood (2010), Franks

et al. (2012)).2The cross-country analysis reveals the following results:

(i) There are large differences in corporate control across legal families.

The share of controlled firms is highest among French civil-law coun-

tries, followed by German and then Scandinavian civil-law countries.

The share of controlled firms is lowest in common-law countries. The

patterns are similar when we look at ownership concentration. These

results, which do not reflect differences in continent, industry,or level of

development, support the early results of the literature (e.g., La Porta

et al. (1999)) using a considerably broader sample of firms and countries.

(ii) Equity blocks are common, present in more than 80% of noncontrolled

firms. This pattern applies across all regions. It also holds in both civil-

law and common-law countries, although the share of widely held firms

with blocks is highest in French civil-law and lowest in common-law

countries.

1The Shapley-Shubik (1954) method is useful for measuring control in firms with multiple large

shareholders and in firms with dispersed ownership and blocks. It also allows one to examine the

precision of the cutoff-based approach employed in extant literature.

2Franks et al. (2012) study corporate control across 4,654 nonfinancial firms in the United King-

dom, France, Germany, and Italy. They find that as firms mature, ownership gets more dispersed

in the United Kingdom, whereas in Italy, Germany, and France family control is higher for older

firms. Foley and Greenwood (2010) document a similar pattern of ownership diffusion in countries

with strong investor protection in a sample of 2,700 firms in 34 countries.