The architecture of attention: Group structure and subsidiary autonomy

• Author: Sharon Belenzon,Niron Hashai,Andrea Patacconi
• Date: 01 October 2019
• DOI: http://doi.org/10.1002/smj.3059
• Published date: 01 October 2019

RESEARCH ARTICLE

The architecture of attention: Group structure

and subsidiary autonomy

Sharon Belenzon

1

| Niron Hashai

2

| Andrea Patacconi

3

1

Fuqua School of Business, Duke University, Durham, North Carolina

2

Arison School of Business The Interdisciplinary Center, Herzliya, Israel

3

Norwich Business School, University of East Anglia, Norwich, UK

Correspondence

Sharon Belenzon, Fuqua School of

Business, Duke University, 100 Fuqua

Drive, Durham, NC 27708-0120.

Email: sharon.belenzon@duke.edu

Abstract

Research Summary:This paper examines the relationship

between strategic decision-making at the subsidiary level

and organizational structure. In many organizations, head-

quarters and subsidiaries are separated by intermediate

subsidiaries. Building on the attention-based view of the

firm, we argue that the greater the “organizational dis-

tance”of a focal subsidiary from headquarters (measured

by the number of intermediate subsidiaries separating the

subsidiary from headquarters), the lower the attention that

headquarters devote to the subsidiary. Thus, subsidiary

autonomy from headquarters increases with organizational

distance. Using a large comprehensive dataset on the struc-

ture of corporate groups in Western Europe, we provide

several pieces of evidence consistent with these hypothe-

ses. By contrast, we find little support for the view that tall

pyramids are created to magnify the voting control of large

shareholders.

Managerial Summary:Corporate groups—confederations

of legally independent firms linked via ownership ties—

are common around the world. An important function of

headquarters in corporate groups is to allocate resources

among member firms. We argue that, because headquar-

ters mostly focus on allocating resources among units that

they directly own, subsidiaries near the top of the group

perform differently in response to changing external

Received: 7 March 2017 Revised: 12 November 2018 Accepted: 29 November 2018 Published on: 5 August 2019

DOI: 10.1002/smj.3059

1610 © 2019 John Wiley & Sons, Ltd. Strat Mgmt J. 2019;40:1610–1643.wileyonlinelibrary.com/journal/smj

conditions than similar unaffiliated firms. This difference

declines as one moves down the group pyramid, as lower-

level affiliates receive less attention from headquarters. An

analysis of a large comprehensive dataset on the structure

of corporate groups in Western Europe supports these pre-

dictions. The paper suggests that the legal organization of

groups is a useful instrument to channel limited headquar-

ters attention to selected affiliates.

KEYWORDS

attention-based view, corporate groups, headquarters, organizational

structure, subsidiary autonomy

1|INTRODUCTION

This paper examines the relationship between strategic decision-making at the subsidiary level and

organizational structure. Recent work on the determinants of subsidiary autonomy has emphasized

factors such as subsidiary capabilities and the balance of power between headquarters and subsidi-

aries (Ambos, Andersson, & Birkinshaw, 2010; Birkinshaw, 1996; Gupta & Govindarajan, 1991;

Martinez & Jarillo, 1989), but has largely neglected structural dimensions of organizational control

emphasized in foundational work (Chandler, 1962; Ethiraj & Levinthal, 2004a; Ghoshal & Bartlett,

1990; Ghoshal & Nohria, 1989; Lawrence & Lorsch, 1967; Stopford & Wells, 1972).

Building on the attention-based view of the firm (Bouquet & Birkinshaw, 2008; Bouquet, Morri-

son, & Birkinshaw, 2009; Joseph & Ocasio, 2012; March & Simon, 1958; Ocasio, 1997), this paper

makes two main contributions: (a) it provides a framework for understanding how managerial atten-

tion is distributed in corporate groups and (b) presents three distinct pieces of evidence consistent

with the proposed framework. Specifically, we propose that “organizational distance”—the number

of intermediate subsidiaries separating a focal subsidiary from headquarters—is a useful construct to

gauge the level of strategic autonomy that the focal subsidiary will enjoy.

Corporate groups—confederations of legally independent firms linked together via ownership

ties—are a prevalent form of organization in both the developed and the developing world

(Belenzon & Berkovitz, 2010; Khanna & Yafeh, 2007; La Porta, Shleifer, & Lopez de Silanes,

1999).

1

To a first approximation, corporate groups exhibit a pyramidal structure, with an “ultimate

owner”(typically a wealthy family, a widely-held corporation, or the state) controlling chains of sub-

sidiary firms. Headquarters perform two fundamental tasks in these structures: they allocate resources

across units, and monitor the performance of these units (Bower, 1970; Collis, Young, & Goold,

2007). Thus, headquarters attention in these structures can be conceptualized at least partly as the

extent to which headquarters are involved in the allocation of budgets and in scrutinizing the

1

For instance, General Electric owns (directly or indirectly) 1,311 subsidiaries (21% domestic) and is, therefore, part of a

corporate group. The 10 largest American industrial corporations own 5,113 subsidiaries, of which 33% are domestic. The

10 largest industrial British corporations own 4,669 subsidiaries (27% are domestic), the 10 largest French corporations own

7,821 subsidiaries (34% are domestic), and the largest 10 industrial German corporations own 5,214 subsidiaries (29%

domestic).

BELENZON ET AL.1611

subsidiaries' strategic plans. The less they are involved, the more the subsidiaries will enjoy auton-

omy, especially when it comes to large investment decisions.

We argue that, as groups expand into related and unrelated businesses, the complexity of manag-

ing a diverse set of businesses will overwhelm the information processing capacity of headquarters

(Chandler, 1962; Ethiraj & Levinthal, 2004a; Simon, 1962). To mitigate this problem, headquarters

will selectively allocate their attention. Because the attention structure of an organization often ech-

oes its formal structure (Joseph & Ocasio, 2012; Joseph & Wilson, 2018), we propose that greater

organizational distance between a focal subsidiary and headquarters will be associated with lower

levels of headquarters attention toward the subsidiary. Thus, organizationally distant subsidiaries will

enjoy greater levels of autonomy than organizationally close subsidiaries.

Subsidiary autonomy may increase with organizational distance (measured by the number of

intermediate affiliates separating the apex firm from the focal subsidiary) because: (a) organizational

distance is inversely related to top management's interest and involvement in the subsidiary and

(b) organizational distance creates communication and governance frictions that reduce the effective-

ness of central control (Bethel & Liebeskind, 1998; Stopford & Wells, 1972). Headquarters may del-

egate parenting responsibilities to intermediate units. However, the mandate of intermediate parents

is likely to be restricted to specific industries and geographies (Stopford & Wells, 1972). Because

subsidiaries near the bottom of the group pyramid are likely to be largely insulated from conditions

in distant parts of the group (e.g., distant industries or geographies), we expect their behavior and

performance to resemble those of similar standalone firms.

Using a large and comprehensive dataset on the internal structure and financial performance of

about 40,000 groups in Western Europe covering the period 2002–2011, we provide several pieces

of evidence consistent with these ideas. First, we show that organizational distance is positively

related to the level of autonomy from headquarters that subsidiary report to enjoy, as measured by

the World Management Survey (Bloom & Van Reenen, 2007). Second, organizational distance is

also related to managerial practices that support decentralized decision-making at the subsidiary

level, such as goal clarity and effective internal control mechanisms. Third, we examine the perfor-

mance implications of greater autonomy. We argue that, if subsidiaries located lower down the group

pyramid are more autonomous from headquarters than subsidiaries located higher up, then their

response to changing industry conditions should be more similar to that of matched standalone firms

than the response of higher-level subsidiaries. All these exercises provide strong support for our

organizational distance hypothesis. The analysis also identifies several factors, including board inter-

locks, family ties and geographical proximity, that reduce but do not eliminate the effects of organi-

zational distance.

By contrast, we find little support for the widespread view that tall pyramids are created to mag-

nify the control of large shareholders (Bebchuk, Kraakman, & Triantis, 2000; Claessens, Djankov, &

Lang, 2000). Among the large groups in our sample, close to 70% are wholly owned or almost

wholly owned by the ultimate shareholder. In the Netherlands and Great Britain this percentage is as

high as 90%. Even in countries, such as Italy, where “pyramiding”is widespread, more than 60% of

all the large groups are wholly owned. Thus, at least in Western Europe and for large groups, the

“control-magnifying”view appears to have limited validity.

Our work has implications for organization theory and corporate restructuring. In a significant

way, the attentional perspective advocated in this paper turns the standard, “control-magnifying”

view of group structure on its head. The control-magnifying view holds that tall structures are created

to magnify the voting control of large shareholders. Our attentional perspective emphasizes instead

1612 BELENZON ET AL.