Managerial Discretion and Performance in China - ReadingSample

Contributions to Management Science

Managerial Discretion and Performance in China

Towards Resolving the Discretion Puzzle for Chinese Companies and Multinationals

Bearbeitet vonHagen Wülferth

1. Auflage 2013. Buch. xxiii, 534 S. HardcoverISBN 978 3 642 35836 4

Format (B x L): 15,5 x 23,5 cmGewicht: 997 g

Wirtschaft > Management > Unternehmensführung

Zu Inhaltsverzeichnis

schnell und portofrei erhältlich bei

Die Online-Fachbuchhandlung beck-shop.de ist spezialisiert auf Fachbücher, insbesondere Recht, Steuern und Wirtschaft.Im Sortiment finden Sie alle Medien (Bücher, Zeitschriften, CDs, eBooks, etc.) aller Verlage. Ergänzt wird das Programmdurch Services wie Neuerscheinungsdienst oder Zusammenstellungen von Büchern zu Sonderpreisen. Der Shop führt mehr

als 8 Millionen Produkte.

http://www.beck-shop.de/Wuelferth-Managerial-Discretion-Performance-China/productview.aspx?product=11891407&utm_source=pdf&utm_medium=clickthru_lp&utm_campaign=pdf_11891407&campaign=pdf/11891407

http://www.beck-shop.de/trefferliste.aspx?toc=9942

http://www.beck-shop.de/fachbuch/inhaltsverzeichnis/9783642358364_TOC_001.pdf

Literature Review and Hypotheses 2

This chapter conducts a thorough, in-depth review of both the empirical and the

theoretical literature on the impact of managerial discretion on performance.

The literature review is used for fulfilling four purposes within the present study:

1. Deriving the research gap and research objective (see Sect. 1.1 and 1.2).

2. Formulating the postulate and hypotheses (see Sects. 2.4.2 and 2.4.3).

3. Selecting the unit of analysis (see Chap. 3).4. Developing the study’s new discretion model (see Chap. 4).

The literature review is structured according to Table 2.1. As indicated by the

rows in Table 2.1, the topics reviewed cover managerial discretion as such (i.e. its

definition and dimensionality), the impact of managerial discretion on performance

(i.e. its direct effect and moderating effects), and the unit of analysis. For each topic,

both the empirical and the theoretical literature are utilised, as represented by the

columns in Table 2.1. The table is completed throughout this chapter in the

following sequence:

• Section 2.1 discusses both the empirical and the theoretical literature for mana-

gerial discretion as such (i.e. its definition and dimensionality).

• Section 2.2 reviews the empirical literature on the impact of managerial discre-

tion on performance, including which units of analysis have been used.

• Section 2.3 reviews the theoretical literature on the impact of managerial

discretion on performance (managerial discretion theory, principal-agent theory,

and stewardship theory), including which units of analysis have been chosen.

• Section 2.4 synthesises the previous sections into implications for the present

study with a focus on formulating the study’s postulate and hypotheses. The other

three purposes are addressed only briefly, as they are further scrutinised in other

chapters (see Chap. 1 for the research gap and research objective, Chap. 3 for the

unit of analysis, and Chap. 4 for the study’s new model).

H. Wulferth, Managerial Discretion and Performance in China,Contributions to Management Science, DOI 10.1007/978-3-642-35837-1_2,# Springer-Verlag Berlin Heidelberg 2013

23

2.1 Managerial Discretion

As noted above, this section discusses both the empirical and the theoretical literature

on managerial discretion, since managerial discretion is the construct that lies at the

heart of the present study. First, Sect. 2.1.1 derives the definition of managerial

discretion from the literature, which forms the basis for measuring discretion in the

study’s empirical model in Chap. 4. Thereafter, Sect. 2.1.2 explores the literature with

respect to the dimensionality of managerial discretion, which again is important for

measuring discretion in the present study’s model and leads to the formulation of the

study’s first hypothesis. Applying the format of Table 2.1, Table 2.2 summarises that

this section reviews the empirical and theoretical literature on managerial discretion,

which generates implications for this study (such as the formulation of Hypothesis 1).

2.1.1 Definition of Managerial Discretion

The present section explores the literature on managerial discretion theory (see

Sect. 2.3.1), principal-agent theory (see Sect. 2.3.2), and stewardship theory (see

Sect. 2.3.3) in an effort to provide a sound definition of managerial discretion. This

forms the basis for measuring discretion in the study’s empirical model in Chap. 4.

The section begins by briefly reviewing the origins of the literature on discretion

before turning to its definition. Discretion is defined in general terms and as applied

to the study’s unit of analysis. This section concludes by explaining why manage-

rial discretion is important—which derives directly from its definition—and by

Table 2.1 Literature review summary (blank)

Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality

Impact on

performance:

• Direct

• Moderating

Unit of analysis

Section 2.1

Section 2.2 Section 2.3

Section 2.4

24 2 Literature Review and Hypotheses

drawing attention to alternative terms that have been used in the literature for

representing the notion of managerial discretion.

The construct of managerial discretion was formally introduced by Hambrick and

Finkelstein (1987) in their managerial discretion theory as the latitude of managerial

action (see below) in order to reconcile polar views of organisational outcomes,

namely the polar views of population ecology (e.g. Aldrich 1979; Baum 1996;

Baum and Amburgey 2002; Carroll 1988; Freeman et al. 1983; Hannan and Freeman

1977, 1984; Singh and Lumsden 1990; Tushman and Romanelli 1985; Zohar and

Luria 2005) and strategic choice theory (e.g. Child 1972, 1997, 2002;Child et al. 2003;

Elbanna and Child 2007; Hitt and Tyler 1991; Hrebiniak and Joyce 1985; Judge and

Zeithaml 1992; Marlin et al. 1994; Miles and Snow 1978; Stienstra et al. 2004; see

Sect. 2.3.1). While studies prior to this seminal work had implicitly assumed the

importance of managerial discretion in determining actions and outcomes of

organisations (e.g. Berle and Means 1932; Chandler 1962) and had at times explicitly

mentioned managerial discretion (e.g. Montanari 1978; Williamson 1963), they had

rarely explicitly defined managerial discretion. After Hambrick and Finkelstein’s

(1987) seminal work, it has continued to be managerial discretion theory (e.g.

Abrahamson and Hambrick 1997; Berman et al. 2005; Carpenter and Golden 1997;

Caza 2011, 2007; Huiyuan Chen 2006; Crossland 2007; Crossland and Hambrick

2007; Datta et al. 2003; Finkelstein and Boyd 1998; Finkelstein and Hambrick 1990;

Finkelstein and Peteraf 2007; Haleblian and Finkelstein 1993; Hambrick and

Abrahamson 1995; Hambrick et al. 1993; Hutzschenreuter and Kleindienst 2007;

Keegan 2006; Keegan and Kabanoff 2008; Key 2002; Yougen Li and Zhao 2004;

Magnan and St-Onge 1997; Quigley and Hambrick 2009; Rajagopalan and

Finkelstein 1992; Thomas and Peyrefitte 1996; Zhang and Li 2008b; Zhang et al.

Table 2.2 Literature review summary (Sect. 2.1 highlighted)

Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality

Impact on

performance:

• Direct

• Moderating

Unit of analysis

Section 2.1


Section 2.4

2.1 Managerial Discretion 25

2006a, b) rather than principal-agent theory (e.g. Agrawal and Knoeber 1996;

Baysinger and Butler 1985; Berger et al. 1997; Brush et al. 2000; Chang and Wong

2003; Childs and Mauer 2008; Denis et al. 1997; Eisenhardt 1989; Fama 1980;

Fama and Jensen 1983a, b; He et al. 2009; Jensen 1986; Jensen and Meckling 1976;

Jensen and Murphy 1990; Jensen and Ruback 1983; Laffont and Martimort 2002;

Lang et al. 1995; Levinthal 1988; Ongore 2011; Shleifer and Vishny 1997;

Spremann 1987; Thepot 2007; Thomsen and Pedersen 2000; Walters 1995;

Wang et al. 2008; Weidenbaum and Jensen 1993; Werner and Tosi 1995, p. 1673;

Xu et al. 2005; Zou 1989) or stewardship theory (e.g. Albanese et al. 1997;Arthurs andBusenitz 2003; Corbetta and Salvato 2004; Davis et al. 1997a, b; Dicke and Ott 2002;

Donaldson 1990; Donaldson andDavis 1989, 1991, 1993, 1994, p. 159; Eddleston and

Kellermanns 2007; Fox and Hamilton 1994; Lane et al. 1999; Liu and Cai 2004;

Miller and Le Breton-Miller 2006; Mills and Keast 2009; Muth and Donaldson 1998;

Salvato 2002; Tian and Lau 2001; Tosi et al. 2003; Van Slyke 2007; Vargas Sanchez

2001, 2004, 2005; Zahra 2003) that has paid the most attention to defining discretion.

Nevertheless, while the definition of discretion in this study therefore derives mainly

from managerial discretion theory, it is consistent with the definitions implicitly used

in principal-agent theory and stewardship theory:

• Contributions to principal-agent theory (e.g. Burkart et al. 1997; Childs and

Mauer 2008; He et al. 2009; Lang et al. 1995) have tended not to explicitly define

managerial discretion, which continues a tendency from new institutional eco-

nomics (Williamson 1963) to pay relatively limited attention to defining the

construct of managerial discretion. This is exemplified by Williamson’s (1963)

paper, which despite carrying the title ‘Managerial Discretion and Business

Behavior’ shies away from explicitly defining discretion. Nevertheless,

Williamson’s (1963) implicit definition of managerial discretion as the latitude

of managers to pursue their own (non-profit-maximising) objectives1 is consis-

tent with the seminal definition by Hambrick and Finkelstein (1987,

pp. 371–378) discussed below. Similarly, more recent studies on principal-

agent theory (e.g. Burkart et al. 1997; Childs and Mauer 2008; He et al. 2009;

Khanchel 2009, p. 97; Lang et al. 1995; Spremann 1987, p. 10) employ the term

managerial discretion—often without explicitly defining it—with an implicit

meaning that concurs with managerial discretion theory.2

1Although Williamson (1963) embeds managerial discretion in an economic theory of the firm

using the notion of expense preference, he does not offer a definition of the term. However, it is

implicit in his work that managerial discretion is viewed as the latitude of managers to pursue their

own (non-profit-maximising) objectives, particularly in terms of channelling the firm’s monopoly

profits to discretionary expenses that benefit the management, such as top management compen-

sation. Williamson (1963) finds some empirical evidence that is consistent with this view, i.e. that

given opportunities for high discretion (e.g. high entry barriers and high internal representation on

the board of directors), discretionary expenses tend to be higher.2 Some studies on principal-agent theory explicitly mention discretion and define it in a way

consistent with Hambrick and Finkelstein’s (1987, pp. 371–378) definition, such as defining

discretion ‘as managers’ decision-making latitude’ (Chang and Wong 2003, p. 2) or as

‘control rights’ (Shleifer and Vishny 1997, p. 742).


• Stewardship theory has likewise placed little emphasis on defining the construct

of managerial discretion and in this regard resembles principal-agent theory

rather than managerial discretion theory. Nevertheless, although stewardship

theory therefore makes no substantial contribution to defining the construct of

discretion, it does mention discretion and uses it in a way consistent with the

definition of managerial discretion provided below (see Sect. 2.3.3; e.g. Davis

et al. 1997b, pp. 25–26; Fox and Hamilton 1994, pp. 70–74; Hambrick and

Finkelstein 1987, pp. 371–378; Mills and Keast 2009, pp. 14–15; Van Slyke

2007, pp. 165–167; Vargas Sanchez 2005, p. 19).

According to the widely-accepted definition by Hambrick and Finkelstein (1987,

pp. 371–378), managerial discretion (or simply discretion)3 is defined as the ‘latitude

of managerial action’, namely the extent to which a manager has multiple courses ofaction (or choices or decisions) across various domains of his/her work that he/sheis aware of and that are acceptable to the parties that possess power to constrain themanager. For instance, the managerial discretion of a plant manager (i.e. the unit of

analysis) measures the extent to which the plant manager has multiple choices across

such domains as making capital investments or hiring workers that he/she is

aware of and that are acceptable to corporate headquarters in China. In particular,

a plant manager with the discretion to undertake small and large capital investments

without prior authorisation from corporate headquarters in China has greater

latitude of action (i.e. discretion) in the domain of making capital investments

than a plant manager who is constrained to making only small capital investments.

As this definition of managerial discretion has been widely accepted in the literature

(e.g. Abrahamson and Hambrick 1997, p. 513; Carpenter and Golden 1997, p. 187;

Caza 2007, p. 27; Chang and Wong 2003, p. 2; Crossland 2007, p. 1; Crossland and

Hambrick 2007, p. 767; Finkelstein and Boyd 1998, p. 179; Finkelstein and

Hambrick 1990, p. 484; Hambrick and Abrahamson 1995, p. 1427; Hambrick and

Finkelstein 1987, pp. 371–378; Rajagopalan and Finkelstein 1992, p. 32), it is

employed in the present study for defining managerial discretion. For this purpose,

the definition of discretion is further scrutinised below in four parts:

• The first part of the definition reads ‘the extent to which a manager has multiplecourses of action (or choices or decisions)’. A manager may have many possible

courses of action within a given domain of his/her work. For instance, within the

domain of making capital investments, a plant manager could potentially choose

between a vast number of alternative investments, such as alternative machinery

or equipment. Although it has been suggested in the literature that one may

attempt to create a complete list of all possible actions that a manager might

take in relevant situations (Hambrick and Finkelstein 1987, p. 401; Yougen Li and

Zhao 2003, pp. 4–5), the approach is generally ruled out on practical grounds,

for even within a given domain, the multiple courses of action available to a

3Managerial discretion is sometimes abbreviated by the term ‘discretion’ in the present study.

The term ‘middle management discretion’ used herein therefore refers to the managerial discretion

of middle management.


manager may be vast and impractical to enumerate (Caza 2007, p. 39; March and

Shapira 1987, p. 1412). The ‘latitude of managerial action’ within a given domain

is therefore commonly specified as the extent to which the manager can autono-

mously decide on his/her courses of action relative to the parties that possess

power to constrain the manager (e.g. Acemoglu et al. 2007; Caza 2007; Chang and

Wong 2003; Cheng et al. 2006; Colombo and Delmastro 2004; Glaister et al.

2003; Marin and Verdier 2006). As described in Sect. 4.2.2, the measurement of

managerial discretion in the present study precisely follows this approach.

• The second part of the definition adds ‘across various domains of his/her work’.A manager can possess multiple courses of action in different areas or domains

of his/her work. While Hambrick and Finkelstein (1987, pp. 371–372) do not

exhaustively postulate an array of domains of managerial action, they provide

examples of domains such as resource allocation, staffing, product market

selection, and competitive initiatives. Similarly, in the present study the domains

of managerial action measured for the plant manager are making capital

investments, hiring workers, introducing new products, and sales and marketing

activities. This selection of domains measured in the present study is shown in

Sect. 4.2.2 to be consistent with the literature (e.g. Acemoglu et al. 2007;

Caza 2007; Chang and Wong 2003; Cheng et al. 2006; Colombo and Delmastro

2004; Glaister et al. 2003; Marin and Verdier 2006).

• The third part refines the definition by specifying ‘that he/she is aware of’.In order for a manager’s multiple courses of action across the domains of his/her

work to count towards his/her level of managerial discretion, a manager must be

aware of these potential choices (Hambrick and Finkelstein 1987, p. 378).

Seminal work has postulated and empirically confirmed that it is thus the

discretion a manager perceives rather than some objective degree of discretion

awarded that matters for predicting managerial behaviour (e.g. Carpenter and

Golden 1997, p. 202; Caza 2007; 2011; Galavan 2005; Galavan et al. 2009;

Glaister et al. 2003; Hambrick and Finkelstein 1987, p. 373; Key 2002;

Walters 1995; Zhao et al. 2010).4 For instance, a plant manager that has been

granted complete authority for hiring a full-time permanent shop floor worker

but erroneously feels that he must involve top management at corporate head-

quarters in the decision process is likely to act subject to this constraint despite a

high degree of objective discretion. As explained in Sect. 4.2.2, the present study

measures discretion based on 467 interviews with plant managers, which allows

the study to evaluate perceived rather than objective managerial discretion of

middle managers.

4 Glaister et al. (2003) find empirical evidence that the managerial discretion a manager perceives

for himself/herself may differ from the discretion that his/her superiors perceive. In particular, in

their sample of UK-European joint ventures, perceptions of managerial discretion of the joint

venture management differ between the joint venture management itself and the parent firms as

well as between each of the parent firms.


• The fourth part of the definition qualifies ‘and that are acceptable to the partiesthat possess power to constrain the manager’. Hambrick and Finkelstein (1987,

p. 378) explain that in order to represent managerial discretion, the multiple

courses of action across the domains that the manager is aware of must ‘lie

within the zone of acceptance of powerful parties’.5 For example, a powerful

party potentially constraining the managerial actions of a chief executive officer

(CEO) is the board of directors (Hambrick and Finkelstein 1987, p. 401),

whereas the powerful party relevant for the plant manager in the present study

is the top management at corporate headquarters in China (see Fig. 1.1 in Sect.

1.3). Hambrick and Finkelstein (1987, p. 401) suggest measuring a CEO’s

managerial discretion as ‘the explicit dollar limits that most firms place on the

CEO’s discretion to commit resources without board approval’ (the board being

the powerful party). In the same sense, the present study measures the plant

manager’s discretion in the domain of making capital investments as the explicit

monetary limit on the maximum capital investment that the plant manager can

undertake without prior authorisation from corporate headquarters in China (i.e.

the relevant powerful party). As demonstrated in Sect. 4.2.2 (see Fig. 4.10), this

study measures the plant manager’s discretion in each of the four domains

assessed relative to the relevant powerful party (i.e. to corporate headquarters

in China). Consequently, the above discussion implies that discretion in the

present study matches each of the four parts of the definition of managerial

discretion.

Summarising the definition, managerial discretion is defined as the extent to

which a manager has multiple courses of action across various domains of his/her

work that he/she is aware of and that are acceptable to the parties that possess power

to constrain the manager (Abrahamson and Hambrick 1997, p. 513; Carpenter and

Golden 1997, p. 187; Caza 2007, p. 27; Chang and Wong 2003, p. 2; Crossland

2007, p. 1; Crossland and Hambrick 2007, p. 767; Finkelstein and Boyd 1998,

p. 179; Finkelstein and Hambrick 1990, p. 484; Hambrick and Abrahamson 1995,

p. 1427; Hambrick and Finkelstein 1987, pp. 371–378; Rajagopalan and Finkelstein

1992, p. 32). By taking the four parts of this definition in turn, it has been shown that

for the unit of analysis of the present study, managerial discretion measures the

5As explained in Sects. 2.3.1 and 2.3.2, Finkelstein and Peteraf (2007, pp. 237–243) incorporate

the assumption of post-contractual asymmetric information (i.e. hidden action) from principal-

agent theory (e.g. Eisenhardt 1989, p. 59; Jensen and Murphy 1990, p. 226; Khanchel 2009, p. 97;

Levinthal 1988, p. 153; Spremann 1987, p. 3; Van Slyke 2007, p. 162; Werner and Tosi 1995,

p. 1673) into managerial discretion theory. They argue that different characteristics of managerial

activities affect the ability of key stakeholders (i.e. the powerful parties) to pre-specify and monitor

the manager’s work, thus creating or constraining discretion. Asymmetric information (i.e. the

inability to monitor the manager’s actions) therefore widens the ‘zone of acceptance of powerful

parties’ (Hambrick and Finkelstein 1987, p. 378). E.g. if top management at corporate headquar-

ters in China could not properly monitor the plant manager’s actions, the plant manager might

undertake capital investments out of self-serving interests that reduced performance, which top

management might have to accept due to their inability to monitor the plant manager’s action

(Spremann 1987, p. 10).


extent to which the plant manager has multiple choices across domains such as

making capital investments, hiring workers, introducing new products, and sales

and marketing activities that he/she is aware of and that are acceptable to corporate

headquarters in China—which is precisely what is measured empirically in the

present study (see Sect. 4.2.2).

The potential importance of managerial discretion in terms of affecting perfor-

mance follows directly from this definition. Defined as the ‘latitude of managerial

action’, managerial discretion measures the leeway of a manager to take action

which might impact on performance (Hambrick and Finkelstein 1987, p. 371).

In fact, as put by Caza (2007, p. 1), ‘[i]ndividuals can only influence organizations

through discretion’ and discretion is thus ‘a fundamental aspect of organized

behavior’ and ‘the key to understanding agency in organizations.’ With the extent

of managerial discretion potentially improving or reducing performance according

to stewardship theory (see Sect. 2.3.3) and principal-agent theory (see Sect. 2.3.2),

respectively, adjusting the discretion granted to middle managers is thus a potential

success factor for the top management of Chinese firms and multinationals in China

when used to optimise the company’s performance along the value chain (e.g.

Adams et al. 2005; Caza 2007; 2011; Chang and Wong 2003; Corbetta and

Salvato 2004; Crossland and Hambrick 2007; Davis et al. 1997b; Donaldson and

Davis 1991; Eddleston and Kellermanns 2007; Finkelstein and Hambrick 1990;

Hambrick and Finkelstein 1987; Hutzschenreuter and Kleindienst 2007; Jensen and

Murphy 1990; Khanchel 2009; Liu and Cai 2004; Mills and Keast 2009;

Misangyi 2002; Quigley and Hambrick 2009; Tang 2008; Tosi et al. 2003;

Van Slyke 2007; Vargas Sanchez 2004; Zhao et al. 2010).6

Finally, it is worth noting that alternative terms have been used in the literature

at times to describe phenomena identical or similar to managerial discretion.

For example, some scholars have employed such terms as managerial autonomyand decision-making autonomy, often synonymously with managerial discretion

(e.g. Cheng et al. 2006; Gammelgaard et al. 2010; Glaister et al. 2003; Groves et al.

1994; Heinecke 2011; Li 2007; Lopez-Navarro and Camison-Zornoza 2003;

Oh 2002; Perrone et al. 2003; Venaik 1999; Verhoest 2003; Wang et al. 2008;

Xu et al. 2005). Other scholars have written about the delegation of authority or

decision rights in terms of decentralisation (e.g. Aghion and Tirole 1997;

Burkart et al. 1997; Colombo and Delmastro 2004; Jensen 1998; Marin and Verdier

2006; Zhang 1997). Certain studies have used autonomy and decentralisation

6 In addition to discretion having a potentially important impact on performance (see above), it has

been empirically demonstrated that discretion may significantly affect managerial power

(Carpenter and Golden 1997), managerial compensation (Finkelstein and Boyd 1998; Magnan and

St-Onge 1997; Rajagopalan and Finkelstein 1992; Werner and Tosi 1995; Wright and Kroll 2002;

Zhang and Xie 2008), workers’ incentives (Groves et al. 1994), a successor chief executive officer’s

age (Wang 2009), top management team tenure, trust (Perrone et al. 2003), strategic attention

(Abrahamson and Hambrick 1997), environmental commitment (Aragon-Correa et al. 2004), pricing

(Cameron 2000), organisational knowledge creation (Oh 2002), and research and development

(Zhang et al. 2006a, b).


interchangeably (e.g. Acemoglu et al. 2007; Bloom et al. 2008), while other studies

have differentiated the two concepts. For example, Barnabas and Mekoth (2010,

pp. 330–336) define autonomy consistently with managerial discretion as the extent

of a manager’s freedom in decision making and decentralisation as the extent to

which this decision-making authority is diffused throughout the organisation. They

therefore argue that autonomy and decentralisation are comparable at lower levels

of operation (e.g. middle management), which they empirically confirm for retail

bank branch managers in India. For the middle managers analysed in the present

study (i.e. plant managers in China; see Chap. 3), managerial discretion, autonomy,

and decentralisation are therefore closely related (e.g. Barnabas and Mekoth 2010,

p. 334; Caza 2007, p. 61), thus allowing the present study to draw on a broader

literature base while keeping given differences in mind.7

2.1.2 Dimensionality of Managerial Discretion

The definition of managerial discretion in the previous section based on Hambrick

and Finkelstein (1987, pp. 371–378) implies that a manager’s discretion

(i.e. ‘latitude of managerial action’) spans various domains of his/her work. In

particular, the domains of managerial action in which a given plant manager can

possess discretion include making capital investments, hiring workers, introducing

new products, and sales and marketing activities in the present study (e.g. Acemoglu

et al. 2007; Caza 2007; Chang and Wong 2003; Cheng et al. 2006; Colombo and

Delmastro 2004; Glaister et al. 2003; Marin and Verdier 2006). However, the

definition of discretion leaves it open whether discretion awarded in one domain

of the manager’s work (e.g. making capital investments) impacts on performance in

a similar way as discretion awarded in another domain of the manager’s work (e.g.

hiring workers)—in which case discretion would be unidimensional—or impacts

on performance in a different way than in different domains—in which case

discretion would be multidimensional. As explained below, whether discretion

can be viewed as unidimensional or multidimensional has important implications

for how to model managerial discretion:

• Unidimensional. If granting the plant manager discretion in one domain of his/

her work (e.g. making capital investments) impacts on performance in a similar

way as granting discretion in any other domain of his/her work (e.g. hiring

workers, introducing new products or sales and marketing activities), then

7Decentralisation of decisions rights from top management at corporate headquarters in China to

the plant manager of the present study is related to the extent to which the plant manager possesses

decision rights (and thus multiple courses of action) across various domains that are acceptable to

top management (i.e. powerful parties). While decentralisation is therefore closely related to

managerial discretion and autonomy in the present study, it tends to emphasise the objective

delegation of decision rights to the plant manager rather than the perceived latitude of managerial

action that defines the managerial discretion of the plant manager.


managerial discretion is unidimensional in its impact on performance. In this

case, measures of a manager’s discretion in different domains can be combined

into a single construct of discretion, which will then, ceteris paribus, rise

whenever discretion increases in any of the measured domains.

• Multidimensional. If, in contrast, granting the plant manager discretion in one

domain of his/her work (e.g. making capital investments) impacts on perfor-

mance in a differently-signed way than granting discretion in another domain of

his/her work, then managerial discretion is multidimensional in its impact on

performance. For example, the plant manager might use discretion for capital

investments in ways that increase performance but use discretion for hiring

workers in ways that decrease performance. If this is this case, then measures

of discretion in different domains cannot meaningfully be combined into a single

(i.e. unidimensional) discretion construct: As a unidimensional construct of

discretion is restricted to impact on performance in a single way (i.e. positive,

neutral or negative), it is not possible for the construct to be increasing in each

discretion measure and still reveal the true, differently-signed impacts of discre-

tion on performance in each domain. Instead, if discretion in different domains

affects performance in distinct ways, then discretion is multidimensional in its

impact on performance and needs to be measured by multiple constructs of

discretion rather than a single, unidimensional construct spanning the various

domains of the manager’s work.

Despite the potential importance of whether discretion is unidimensional or

multidimensional (which motivates the present study’s postulate; see Sects. 1.2

and 2.4), the theoretical and empirical literature to date have not provided a

conclusive answer as to whether discretion should be treated as unidimensional or

as multidimensional (i.e. whether or not managerial discretion has distinct impacts

in different domains). In particular, there seems to be a discontinuity between

managerial discretion theory on the one hand, which postulates the multidimen-

sionality of discretion, and many empirical studies as well as principal-agent theory

and stewardship theory on the other hand, which implicitly tend to assume the

unidimensionality of discretion:8

• On the one hand, both early and recent theoretical studies grounded mostly in

managerial discretion theory (see Sect. 2.3.1) have posited on qualitative groundsthat managerial discretion consists of several types, i.e. is multidimensional (e.g.

Barnabas and Mekoth 2010; Carpenter and Golden 1997, p. 195; Caza 2007, pp.

26–82; Chen 2006; Finkelstein and Peteraf 2007, p. 245; Groves et al. 1994, p. 190;

Hambrick and Abrahamson 1995, p. 1439; Hambrick and Finkelstein 1987, pp.

371–402;Hambrick et al. 1993, p. 409;March and Simon 1958; Perrone et al. 2003,

pp. 422–423). In particular, Hambrick and Finkelstein (1987, pp. 371–402) contend

that managers vary significantly ‘in the number of domains in which they have

8 This discontinuity motivates the study’s first hypothesis (see below) and the integration of the

various existing theories into a new single model for the impact of managerial discretion on

performance (see Chap. 4).


discretion’ and thus ‘have some domains of high discretion and others of low

discretion’. They expect that this ‘type of mixed discretion lead[s] to consequences

that differ from simply thinking about moderate discretion’ (1987, p. 402). More-

over, they posit that each different combination of causes of discretion ‘may lead to

its own set of accompanying organizational factors’, including performance (1987,

p. 389). Applied to the unit of analysis of the present study, managerial discretion

theory therefore implies that a plant manager might have different degrees of

discretion in different domains of his/her work, and a plant manager with high

discretion for making capital investments and low discretion for hiring workers

might produce different results than a plant manager with moderate discretion for

bothmaking capital investments and hiringworkers. In other words, it is implicit in

managerial discretion theory that discretion in different domains might lead to

distinct impacts and in this sense might be a multidimensional construct.

• On the other hand, as explained in the preceding section, neither principal-agenttheory (see Sect. 2.3.2) nor stewardship theory (see Sect. 2.3.3) have placed

much emphasis on the nature of the construct of discretion. When referring to

discretion, both theories tend to treat discretion as a single (unidimensional)

construct with a particular positive (in stewardship theory) or negative (in

principal-agent theory) impact on performance (e.g. Chang and Wong 2003;

Dicke and Ott 2002, p. 468; Fox and Hamilton 1994, p. 74; He et al. 2009;

Spremann 1987, p. 18; Vargas Sanchez 2005, p. 19; Xu et al. 2005). Moreover,

despite the postulated multidimensionality in qualitative studies of managerial

discretion theory (see above), quantitative empirical studies have frequently

made the simplifying assumption in their research designs that discretion is

unidimensional. In particular, among studies attempting to measure discretion

directly, scholars have frequently gauged a manager’s discretion in different

areas of his/her work and combined these indicators into a single unidimensional

discretion construct (e.g. Barnabas and Mekoth 2010; Bloom et al. 2008;

Caza 2007; 2011; Chang and Wong 2003; Cheng et al. 2006; Gammelgaard

et al. 2010; Marin and Verdier 2006). Moreover, among the many empirical

studies resorting to proxy measures for gauging managerial discretion, discretion

has prevalently been modelled as unidimensional as well.9 Compared to this

prevalent assumption of unidimensionality in the empirical literature, only a

9 Empirical studies have modelled unidimensional discretion constructs by measuring one or

several proxies related to e.g. ratings of managerial power, internal representation on the board

of directors, managerial stock ownership, and financial ratios (e.g. Huiyuan Chen 2006;

Khanchel 2009; Yougen Li and Zhao 2004; Zhang and Li 2008b; Zhang et al. 2006a, b) as well

as multiple antecedents drawn from mostly the task environment (e.g. Agarwal et al. 2009;

Berman et al. 2005; Cameron 2000; Finkelstein and Boyd 1998; Finkelstein and Hambrick

1990; He et al. 2009; Magnan and St-Onge 1997; Rajagopalan and Finkelstein 1992;

Williamson 1963). In addition, industry-level discretion has been frequently proxied in existing

studies (e.g. Abrahamson and Hambrick 1997; Datta et al. 2003; Finkelstein and Hambrick 1990;

Hambrick and Abrahamson 1995; Hambrick et al. 1993; Keegan 2006; Keegan and Kabanoff

2008; Thomas and Peyrefitte 1996).


small minority of studies have differentiated managerial discretion according to

multiple dimensions of the manager’s work and these studies have often

analysed phenomena other than the impact of managerial discretion on perfor-

mance (e.g. Colombo and Delmastro 2004; Glaister et al. 2003; Groves et al.

1994; Xiaoyang Li 2007; Perrone et al. 2003).

Despite this discontinuity of postulated multidimensionality (in managerial

discretion theory) and implicitly assumed unidimensionality (in many empirical

studies as well as principal-agent theory and stewardship theory), empirical studies

have on occasion investigated the dimensionality of discretion explicitly. This has,

however, not always produced concurrent results and therefore warrants further

investigation. For example, Cheng et al. (2006, p. 348) experiment with separating

operational and strategic decisions but find that these two indices are sufficiently

correlated and correlate sufficiently with the overall index in order to support a

unidimensional measure for discretion. By contrast, Caza’s (2007, pp. 26–82)

factor analysis confirms that research and development (R&D) managers in Europe

perceive their discretion as multidimensional and his regression reveals that some

antecedents (or causes, determinants or sources of discretion; see Sect. 2.3.1) differ

between dimensions of discretion. Furthermore, Caza generalises the multidimen-

sionality of discretion on qualitative grounds by demonstrating its consistency with

the literature on various hierarchical levels in organisations from workers to top

management. However, when empirically investigating the impact of discretion on

performance, Caza (2007, pp. 14–16) does not allow for this multidimensionality,

implicitly treating discretion as unidimensional and thus modelling only a single

overall impact on performance.

The unresolved dimensionality of managerial discretion in the theoretical and

empirical literature (i.e. unidimensionality versus multidimensionality) motivates

the formulation of one of the four hypotheses of the present study (i.e. Hypothesis

1). As discussed above, although managerial discretion theory hints that discretion

may be multidimensional, both principal-agent theory and stewardship theory tend

to treat discretion as unidimensional and many existing empirical studies have

tended to make the simplifying assumption in their research designs that discretion

is unidimensional (see references above). Whether or not this assumption of

unidimensionality is universally tenable is examined by testing the null hypothesis

of unidimensionality (H01) against the alternative hypothesis of multidimensionality

(H11)—i.e. testing whether the impact of discretion on performance differs between

dimensions of discretion. If H01 cannot be rejected, the prevalent simplifying

assumption of a unidimensional construct of discretion in existing studies might

be justifiable. However, if H01 can be rejected in favour of H1

1 , this simplifying

assumption of unidimensionality is not universally tenable (since it is then not

tenable at least in the instance of this particular study) and instead more granular

research designs that allow for the potentially multidimensional nature of discretion

would be required in order to produce more meaningful results in future research.


Hypothesis 1 (Dimensions of Discretion)

H01: Managerial discretion is unidimensional in its impact on performance.

H11: Managerial discretion is multidimensional in its impact on performance.

The importance of testing Hypothesis 1 in the present study derives from its

potential contribution towards resolving the discretion puzzle and thus towards

fulfilling the present study’s research objective (see Sect. 1.2). In particular, as is

explained in Chap. 7, erroneously treating discretion as unidimensional when it is in

fact multidimensional might produce misleading estimates of the impact of discre-

tion on performance that could potentially be a cause of the contradictory empirical

evidence that gives rise to the discretion puzzle. Therefore, empirically testing

Hypothesis 1 regarding the dimensions of managerial discretion in the present study

constitutes a vital step towards resolving the discretion puzzle.

Table 2.3 summarises the discussion on the definition and dimensionality of

discretion presented in this section (i.e. Sect. 2.1) using the literature review sum-

mary table introduced in the opening of Chap. 2 (see Table 2.1). As indicated by the

grey-shaded arrow in Table 2.3, the above discussion of the theoretical and empirical

literature leads to the formulation of Hypothesis 1 on the dimensionality of discre-

tion. The following sections complete the remaining cells in this table, which allows

Sect. 2.4 to eventually synthesise the theoretical and empirical literature into

implications for the present study, such as the study’s four research hypotheses.

2.2 Impact of Managerial Discretion on Performancein Existing Empirical Evidence

This section conducts a thorough, in-depth review of empirical studies10 on the

impact of managerial discretion on performance, which completes the grey-shaded

10 The present study reviews over 80 empirical studies on managerial discretion and related

phenomena, e.g. Abrahamson and Hambrick (1997), Acemoglu et al. (2007), Adams et al. (2005),

Agarwal et al. (2009), Agrawal and Knoeber (1996), Aragon-Correa et al. (2004), Barnabas and

Mekoth (2010), Baysinger and Butler (1985), Berger et al. (1997), Berman et al. (2005), Bloom et al.

(2008), Bowen et al. (2008), Brush et al. (2000), Burkart et al. (1997), Zhang and Li (2008b), Zhang

and Xie (2008), Zhang et al. (2006a, b), Cameron (2000), Carpenter and Golden (1997), Caza

(2007), Caza (2011), Chaganti et al. (1985), Chang and Wong (2003), Chang and Wong (2004),

Chen (2006), Cheng et al. (2006), Colombo and Delmastro (2004), Crossland and Hambrick (2007),

Datta et al. (2003), Demsetz and Lehn (1985), Denis and Denis (1993), Denis et al. (1997),

Donaldson and Davis (1991), Finkelstein and Boyd (1998), Finkelstein and Hambrick (1990),

Gammelgaard et al. (2010), Glaister et al. (2003), Groves et al. (1994), Wang (2009), Haleblian

and Finkelstein (1993), Hambrick and Abrahamson (1995), Hambrick et al. (1993), He et al. (2009),

Heinecke (2011), Hutzschenreuter and Kleindienst (2007), Kayhan (2008), Keegan and Kabanoff

(2008), Keegan (2006), Kesner (1987), Khanchel (2009), Lang et al. (1995), Lieberson and

O’Connor (1972), Lopez-Navarro and Camison-Zornoza (2003), Mackey (2008), Magnan and St-

Onge (1997), Manner (2010), Marin and Verdier (2006), Misangyi (2002), Oh (2002), Ongore

2.2 Impact of Managerial Discretion on Performance in Existing Empirical Evidence 35

cells in the literature review summary Table 2.4.11 To this end, four steps are

pursued:

1. Section 2.2.1 conceptually decomposes each empirical study into two parts,

namely its research design and its empirical results, which paves the way for

scrutinising the existing empirical evidence in the remainder of this section.

2. Section 2.2.2 presents the differences in empirical results between existing

studies that give rise to the discretion puzzle (see Sect. 1.1), i.e. the positive,

neutral, and negative estimated impacts of discretion on performance.

Table 2.3 Literature review summary (Sect. 2.1 completed)

Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality

• The extent to which a manager has multiple courses of action across

various domains of his/her work that he/she is aware of and that are

acceptable to the parties that possess power to constrain the manager

• Discontinuity between postulated multidimensionality in managerial

discretion theory and assumed unidimensionality of discretion in

empirical evidence, principal-agent theory, and stewardship theory

Impact on

performance:

• Direct

• Moderating

Unit of analysis


Section 2.4

Hypothesis 1

(2011), Palmer (1973), Perrone et al. (2003), Quigley and Hambrick (2009), Rajagopalan and

Finkelstein (1992), Stano (1976), Wang et al. (2008), Tang (2008), Thomas and Peyrefitte (1996),

Venaik (1999), Verhoest (2003), Zhang (1997), Walters (1995), Werner and Tosi (1995),

Williamson (1963), Wright and Kroll (2002), Li (2007), Xu et al. (2005), Li and Zhao (2004),

Yan et al. (2010), Zahra and Stanton (1988), Zhao et al. (2010), and Zheng (2007).11 As explained below, the impact of managerial discretion on performance can be decomposed

into a direct effect and moderating effects, which are denoted by ‘Direct’ and ‘Moderating’ in

Table 2.4, respectively. Moreover, as the review of the literature on the impact of discretion on

performance has implications for the choice of the unit of analysis of the present study, Table 2.4

includes the ‘Unit of analysis’ as an additional row.


3. Section 2.2.3 explores the differences in research designs between existing

studies so as to diagnose potential causes for why their estimated impacts of

discretion on performance might have differed from each other across studies.

4. Section 2.2.4 derives implications for the hypotheses of the present study, i.e. itformulates hypotheses for testing potential causes derived from the differences

between empirical studies in Sect. 2.2.3 within the present study. These

hypotheses are entered into the literature review summary Table 2.4, as denoted

by the grey-shaded arrow therein.

2.2.1 Conceptual Decomposition of Empirical Studies

In order to present the existing empirical evidence in a comparable manner, this

section conceptually decomposes each empirical study into two parts, namely its

research design and its empirical results. The term research design here refers to allthe methodological aspects in the empirical study that produce the empirical resultson the impact of managerial discretion on performance.12 A research design

includes:


Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality







Impact on

performance:

• Direct

• Moderating

Unit of analysis


Section 2.4

Hypothesis 1

12 Similar research design definitions can be found in the literature (e.g. Chui 2007, p. 66;

Punch 2005, p. 62).


• A unit of analysis, which is ‘the entity about which one is trying to draw

conclusions’ (Johnson et al. 2007, p. 58) and in the studies reviewed below

generally refers to a manager whose managerial discretion is investigated

(e.g. top management in the United States).13 Most studies’ results are estimated

based on a sample drawn from the population that the unit of analysis defines

(Northrop and Arsenault 2007, p. 214).

• A measure of discretion and a measure of performance, which attempt to gauge

discretion and performance for the study’s chosen unit of analysis.14

• A modelling methodology for estimating the impact of the measure of discretion

on the measure of performance for the chosen unit of analysis. The modelling

methodology can be decomposed into the equations that are estimated for

inferring the impact of discretion on performance and the multivariate analysis

technique that is applied for estimating these equations.

– As to multivariate analysis techniques, existing studies typically adopt obser-vational cross-sectional designs (e.g. Caza 2007; 2011; Huiyuan Chen 2006;

Colombo and Delmastro 2004; Finkelstein and Boyd 1998; Glaister et al.

2003; Yougen Li and Zhao 2004; Marin and Verdier 2006; Oh 2002; Perrone

et al. 2003; Williamson 1963; Zhang and Li 2008b; Zhang et al. 2006a, b)

using first and second generation multivariate analyses, namely multiple

regressions (e.g. ordinary least squares multiple regressions) and structural

equation models (i.e. variance-based and covariance-based structural equa-

tion models; see Sect. 5.1.1; e.g. Bentler and Chou 1987; Bentler and Weeks

1980; Bollen 1989; Chin et al. 2003, p. 194; Fassott 2005; Fornell 1987;

Joreskog 1970, 1981; Joreskog and Sorbom 1982, 1988; Lohmoller 1987,

1989; Wold 1966, 1973, 1975, 1982, 1985, 1989).15

– In order to estimate the impact of managerial discretion on performance, the

chosen multivariate analysis technique in a given study then typically

estimates an equation that resembles Equation (2.1):16

13 The unit of analysis of the present study (i.e. the plant manager in China) is discussed in detail in

Chap. 3.14 The measures of discretion and performance in the present study are described in Sect. 4.2 and are

demonstrated to exhibit high reliability and construct validity in Sects. 5.2 and 5.3, respectively.15 A discussion of alternative multivariate analysis techniques is provided in Sect. 5.1.1, which

develops its own decision-tree logic for choosing an appropriate modelling methodology for the

present study.16 A more complete version of Equation (2.1) is developed from the literature in Sect. 4.3 as part of

the present study’s new empirical discretion model. This equation resembles Equation (2.1) when

expressed in vector/matrix notation (e.g. Gentle 2007, pp. 479–491; Harville 2008, pp. 1–10;

Knapp 2007, pp. xxi–xxiv), but it is sufficiently general that it can be disaggregated into an

arbitrary number of discretion dimensions, controls, and moderators. In contrast to the simplified

Equation (2.1), the equation in Sect. 4.3 fulfils the requirements of state-of-the-art methodological

research such as that all the components of the product term D �Mð Þ must be included in the

equation in direct form (Carte and Russell 2003, pp. 480–495; Cohen 1978; Cronbach 1987;

Henseler and Fassott 2010, pp. 718–719; Irwin and McClelland 2001, p. 105). Applying this

requirement to Equation (2.1), it would read P ¼ d � Dþ c � Cþ m � D �Mð Þ þ q �M þ ε . Theintercept term is normalised to zero.


P ¼ d � Dþ c � Cþ m � D �Mð Þ þ ε; (2.1)

where P is the measure of performance and D the measure of discretion (see

above), C is a control variable to remedy spurious effects (e.g. Jaccard and

Turrisi 2003, pp. 1–2; Simon 1954, pp. 477–478), M is a moderator variable

(e.g. Baron and Kenny 1986, p. 1174; Henseler and Fassott 2010, p. 714), and

ε is an error term (e.g. Betzin and Henseler 2005, p. 53; Gentle 2007, p. 490).

The lower-case letters are the parameters in a study’s equation that are

calibrated by applying the study’s chosen multivariate analysis technique to

the study’s sample of its unit of analysis (see above). In particular, d estimates

the direct effect of discretion (D) on performance (P), c estimates the control

effect of the control variable (C) on performance (P), and m estimates the

moderating effect of the moderator variable (M) on the impact of discretion

(D) on performance (P).17 Studies differ in terms of whether and how many

control variables (C) and moderator variables (M) are included, so a study

with two controls and no moderators would, for instance, estimate an equa-

tion such as P ¼ d � Dþ c1 � C1 þ c2 � C2 þ ε .18 Nevertheless, in principle

the equations tend to follow the structure of Equation (2.1), which makes

Equation (2.1) a suitable simplification for presenting existing empirical

evidence in a comparable manner.

In addition to possessing a research design (i.e. unit of analysis, measures of

discretion and performance, and modelling methodology), an empirical study was

characterised above as consisting of empirical results on the impact of managerial

discretion on performance, which are produced by applying the research design.

Specifically, once an empirical study has collected a sample of the unit of analysis

with measures of discretion and performance, it can apply its modelling methodol-

ogy so as to obtain sample estimates for the parameters in Equation (2.1) (i.e. d, c,and m). Moreover, provided the study’s sample is statistically representative of the

relevant population of the unit of analysis, the empirical study can test for the

17A discussion of direct effects, control effects, mediating effects, and moderating effects is

presented in Sect. 4.3. Moreover, a discussion of control variables is included in Sect. 5.4.1 on

the threat to internal validity of excluding associated variables.18While there are differences in the equations estimated across existing studies other than in the

number of controls (C) and moderators (M), the overall structure of the equations remains similar to

Equation (2.1). Some studies add such terms to Equation (2.1) as intercept terms [which are

normalised to zero in Equation (2.1)], quadratic terms so as to model non-linear effects

(e.g. Huiyuan Chen 2006; Zhang and Li 2008b, p. 122) or lagged terms when time series are

involved (e.g. Cheng et al. 2006; Groves et al. 1994). Likewise, instead of modelling moderating

effects with the product term approach in Equation (2.1) (see Sect. 4.3.1; e.g. Carte and Russell

2003, pp. 480–495; Chin et al. 2003, pp. 196–200; Irwin and McClelland 2001, p. 105), certain

studies exclude the product term D �Mð Þ and apply the multi-group comparison approach for

testing moderating effects (e.g. Arnold 1982; Henseler and Fassott 2010, pp. 719–721; Rigdon

et al. 1998, p. 1; Venkatraman 1989, p. 426). Yet even these studies can be conceptually

represented by Equation (2.1), since these two approaches to modelling moderating effects

resemble each other when the moderator can be sensibly dichotomised (Henseler and Fassott

2010, p. 721; Qureshi and Compeau 2009, p. 199).


significance of the parameters in Equation (2.1) (see Sects. 4.3.3 and 5.5; e.g. Fogiel

2000, pp. 158–190; Garson 2002, pp. 139–196; Gliner and Morgan 2000, p. 148;

Greene 2003, pp. 892–896; Gujarati 2004, pp. 119–139; Hayashi 2000, pp. 33–45;

Salvatore and Reagle 2002, pp. 87–95; Spanos 1986, pp. 213–311; Wooldridge

2002, pp. 116–299).

The way in which these empirical results on the parameters in Equation (2.1)

(i.e. d , c , and m) relate to the impact of managerial discretion on performance

can be seen by partially differentiating performance (P ) in Equation (2.1) with

respect to discretion (D) so as to obtain an equation of discretion’s total impact

on performance:

@P

@D¼ d þ m �M (2.2)

Equation (2.2) reveals that the total estimated ceteris paribus impact of manage-

rial discretion (D) on performance (P) is the sum of the direct effect of discretion on

performance (here d) and the moderating effect of the moderator variable (here m)multiplied by the value of the moderator variable (M). If variables are centred to

means of zero, then when the moderator variable reaches its mean value (i.e. zero),

the total impact of discretion on performance measures the direct effect of discre-

tion on performance (see Sect. 4.3.2 on comparative statics; e.g. Aiken and West

1991, p. 37; Dowling 2000, pp. 284–291; Finney et al. 1984; Henseler and Fassott

2010, p. 728; Hirschey 2009, p. 99). Consequently, Equation (2.2) implies that the

total impact of discretion on performance is equal to the direct effect of discretion

on performance (d) when the moderator variable is zero (e.g. at its average level)

and is adjusted upwards or downwards to the extent that the moderator diverges

from zero. For example, if the moderator measures firm size, with M ¼ 0

representing an average-sized firm and M ¼ 1 a large firm, then the total impact

of discretion on performance is d þ m for a large firm, i.e. it exceeds the impact of

discretion in an average-sized firm (d) by the moderating effect of firm size (m).Given that the existing empirical studies on the impact of discretion on perfor-

mance have differed from each other as to whether and which moderators were

included (with most studies not modelling any moderators; see Sect. 2.2.4), only the

direct effects of discretion on performance (d) are fully comparable across studies.

Hence, when setting moderators, where included, to zero (i.e. settingM ¼ 0, which

reduces Equation (2.2) to d ), the total impact of discretion on performance in

existing studies can be inferred from the estimated direct effect of discretion on

performance (d):• If the direct effect is significantly greater than zero (d > 0 ), then discretion

is estimated as having a positive impact on performance, meaning granting a

manager more managerial discretion might increase performance.

• If the direct effect is significantly smaller than zero (d < 0), then discretion is

estimated as having a negative impact on performance, meaning granting a

manager more managerial discretion might reduce performance.


• If the direct effect is insignificantly different from zero (d � 0), then discretion is

found to have an insignificant or neutral impact on performance, meaning

granting a manager more managerial discretion might not alter performance.

It should be noted, however, that an insignificant direct effect (d � 0) does not

prove discretion to have no impact on performance, since there might be a non-

zero impact on performance in the population, but e.g. the study’s sample might

be too small to provide significant evidence (Betton 1985, p. 3; Doehring 1988,

p. 104; Zhuravskaya 2000, p. 143).

Having conceptually decomposed empirical studies into their research designs

and empirical results, the next two sections review the extant literature with respect

to differences in empirical results (e.g. positive versus neutral versus negative

impacts on performance, d ; see Sect. 2.2.2) and differences in research designs

(e.g. unit of analysis, measures of discretion and performance, and modelling

methodology; see Sect. 2.2.3) in an effort to work towards resolving the discretion

puzzle.

2.2.2 Differences in Empirical Results

Drawing on the methodological discussion in the previous section (see Sect. 2.2.1),

this section compares and contrasts the empirical results on the impact of discretion

on performance in the extant literature in terms of the studies’ estimated directeffects of discretion on performance ( d in Equations (2.1) and (2.2) above).19

Consistently finding a positive (d > 0), neutral (d � 0) or negative (d < 0) direct

effect throughout existing studies would demonstrate that managerial discretion

tended to improve, not alter or reduce performance, respectively. However, the

review reveals that there is abundant contradictory evidence on discretion’s

impact on performance, ranging from positive (d > 0) to neutral (d � 0) and

negative (d < 0). It is therefore unclear whether additional discretion can be

expected to improve, not alter or reduce performance and this gives rise to the

discretion puzzle discussed in Sect. 1.1:20

• A number of scholars have found significant evidence that the direct effect of

discretion on performance is positive (d > 0), i.e. that granting a manager more

managerial discretion might increase performance (e.g. Agarwal et al. 2009;

Barnabas and Mekoth 2010; Chang and Wong 2003; Gammelgaard et al. 2010;

Khanchel 2009). These references demonstrate that empirical evidence of a

positive impact of discretion on performance spans different levels of

19 As explained in Sect. 2.2.1, when moderators deviate from zero, the impact of discretion on

performance deviates from the direct effect of discretion (d ) by moderating effects, which are

reviewed in Sect. 2.2.4.20 In order to achieve a broad coverage of existing empirical evidence, studies are also included if

they denote managerial discretion by alternative comparable terms, such as managerial autonomy

(see Sect. 2.1.1).


management and geographies, e.g. hedge fund managers worldwide (Agarwal

et al. 2009), retail bank branch managers in India (Barnabas and Mekoth 2010),

top management in China (Chang and Wong 2003), management of foreign-

owned subsidiaries of multinationals in Europe (Gammelgaard et al. 2010), and

top management in Tunisia (Khanchel 2009).

• However, various scholars have also found significant evidence that the direct

effect of discretion on performance is negative (d < 0), i.e. that granting a

manager more managerial discretion might reduce performance (e.g. He et al.

2009; Heinecke 2011; Stano 1976; Williamson 1963; Xu et al. 2005).

Again, these studies exemplify that discretion might harm performance for

various levels of management and geographies, e.g. top management in the

United States (He et al. 2009; Stano 1976; Williamson 1963), regional manage-

ment of multinationals worldwide (Heinecke 2011), and top management in

China (Xu et al. 2005).

• Finally, scholars have found the direct effect of discretion on performance to be

insignificant (d � 0), meaning granting a manager more managerial discretion

might not alter performance (e.g. Caza 2011; Groves et al. 1994; Yougen Li and

Zhao 2004; Lopez-Navarro and Camison-Zornoza 2003; Venaik 1999). This

evidence on discretion not significantly altering performance likewise covers

multiple levels of management and geographies, e.g. managers of research and

development (R&D) units in Europe (Caza 2011), factory managers in China

(Groves et al. 1994), top management in China (Yougen Li and Zhao 2004),

general managers of Spanish export joint ventures (Lopez-Navarro and

Camison-Zornoza 2003), and marketing managers of subsidiaries of

multinationals worldwide (Venaik 1999).

While this contradictory evidence21 on the impact of discretion therefore spans

different levels of management and different geographies, it should be noted that

even for a given level of management in a given country, empirical studies have

estimated positive, neutral, and negative impacts of discretion on performance. This

is exemplified by top management in China, where Chang and Wong (2003, 2004)

and Zhang (1997) find a positive, Li and Zhao (2004) find a neutral, and Xu et al.

(2005) find a negative direct effect of managerial discretion on performance.

In short, there is an abundance of coexisting empirical evidence that managerial

discretion has a positive, neutral, and even negative impact on performance.

21 It should further be noted that in addition to empirical studies that explicitly aim to measure

managerial discretion or autonomy (exemplified by the references above), the contradictory

evidence extends to studies that implicitly measure constructs potentially related to discretion,

such as diffusion of ownership, managerial stock ownership, internal representation on the board,

and other measures of board composition. Chang and Wong (2003, p. 7) view such studies as

evidence of an inconclusive relationship between discretion and performance, with the empirical

findings of e.g. Donaldson and Davis (1991) and Kesner (1987) supporting a positive, those of

Chaganti et al. (1985), Demsetz and Lehn (1985), and Zahra and Stanton (1988) supporting a

neutral, and those of Baysinger and Butler (1985) and Palmer (1973) supporting a negative

relationship.


Section 1.1 coined the term discretion puzzle for this ostensible paradox that accordingto empirical evidence, discretion increases, does not affect, and decreases performance,

which remains unexplained by the existing theories (i.e. managerial discretion theory,

principal-agent theory, and stewardship theory; see Sects. 1.1 and 2.3).

A graphical representation of this contradictory (i.e. positive, neutral, and nega-

tive) empirical evidence on the impact of managerial discretion on performance

is given by the fictitious scatter plot in Fig. 2.1.22 The horizontal axis measures

the extent of the manager’s discretion (D), ranging from low managerial discretion

(left-hand side) to high managerial discretion (right-hand side), while the vertical

axis measures the resulting performance (P), ranging from low performance (bottom)

to high performance (top). The dotted lines depict the estimated function of

performance as a function of discretion from Equation (2.1) above (i.e. P ¼ d � Dþ c � Cþ m � D �Mð Þ) for the case where potential control variables and moderator

variables are equal to zero (i.e. C ¼ M ¼ 0), namely P ¼ d � D .23 The upwards-

sloping dotted line therefore represents empirical evidence which finds that additional

discretion tends to improve performance d > 0). Likewise, the downwards-sloping

line denotes empirical evidence that finds a negative impact of discretion on

High Performance (P)

LowManagerialDiscretion (D)

HighManagerial

Discretion (D)

Low Performance (P)

DiscretionPuzzle

Neutral ( d ≈ 0) (e.g.Caza, 2011; Groves, et al., 1994; Y. Li & Zhao, 2004; López-

Navarro & Camisón-Zornoza, 2003;Venaik, 1999)

Fig. 2.1 Fictitious scatter plot for discretion puzzle

Source: Selection of empirical studies on the impact of managerial discretion on performance

22 The scatter plot format in Fig. 2.1 is applied in Sects. 2.3 and 2.4 below in order to link the

theories to the empirical evidence on the impact of discretion on performance and derive the

present study’s Hypothesis 4.23 Non-zero values of a control variable (C 6¼ 0) simply shift the dotted lines in Fig. 2.1 upwards or

downwards (provided there is a non-zero control effect, i.e. c 6¼ 0) without affecting the slopes of

the lines. By contrast, non-zero values of a moderator variable (M 6¼ 0) alter the slope and therefore

tilt the dotted lines (if m 6¼ 0).


performance (d < 0). Finally, the flat line depicts empirical evidence that finds an

insignificant (or neutral) impact of discretion on performance (d � 0). This fictitious

scatter plot therefore reveals that when treating existing empirical evidence at an

aggregate level in terms of their estimated impact of managerial discretion on

performance, their results are divergent, ranging from positive to neutral and even

negative. These contradictory results give rise to the discretion puzzle, as indicated in

Fig. 2.1.24 In order to work towards resolving the discretion puzzle, the next section

explores differences in the research designs of existing studies as potential causes for

their widely differing estimated impacts of discretion on performance.

2.2.3 Differences in Research Designs

In order to work towards resolving the discretion puzzle (i.e. the research objective;

see Sect. 1.2), this section turns to diagnosing potential causes for why existing

empirical studies might have produced such divergent empirical results as shown in

Fig. 2.1 in Sect. 2.2.2. As Sect. 2.2.1 has defined a study’s research design as all

methodological aspects that produce a study’s particular empirical results, differences

in research designs between existing studies could be potential causes for differences

in empirical results between existing studies (i.e. differences in d). In particular, the

differences in the chosen unit of analysis, measures of discretion and performance,

and modelling methodology between existing studies might be able to explain the

positive, neutral, and negative estimated impacts of managerial discretion on perfor-

mance across studies and thereby help resolve the discretion puzzle.25

Yet it is first of all necessary to identify the differences in aspects of the research

designs underlying the existing studies under review before maintaining that such

aspects can at least in part explain disparities in the empirical results of those

studies. This section therefore explores whether and which research design aspects

have differed between studies, thus diagnosing whether they could be potential

causes for the observed differences in empirical results. It is found that empirical

studies have indeed varied strongly in their research designs, enforcing the notion

that the differences in research designs might help explain the differences in the

estimated results for the impact of discretion (i.e. positive, neutral, and negative):

24 It should be noted that due to their observational cross-sectional designs, the existing empirical studies

cannot unequivocally demonstrate that additional discretion causes an increase or decrease in perfor-

mance but rather only make statements regarding association that may be consistent with causality

(e.g. Caza 2007, p. 46; Finkelstein and Hambrick 1990, p. 500; Granger 1969; Sanchez 2008, p. 5;

Simon 1954, pp. 477–478; Wagner 2002, pp. 287–292; see Sect. 5.4.1).25 For example, if two different studies adopt two different measures of discretion, each of which

taps into a distinct dimension of discretion, then one study might find a positive and the other a

negative impact of discretion on performance, provided the distinct discretion dimensions have

different performance impacts.


• The unit of analysis (i.e. manager whose discretion is analysed) has varied across

studies on the impact of discretion in a number of ways, including with respect to

geography, level of management, firm type, and firm size:

– Geography. Empirical studies have analysedmanagers worldwide (e.g. Agarwal

et al. 2009; Crossland and Hambrick 2007; Venaik 1999), in the United States

(e.g. He et al. 2009; Stano 1976; Tang 2008), in European countries (e.g. Caza

2011; Gammelgaard et al. 2010), in Asian countries (e.g. Barnabas and Mekoth

2010; Yougen Li and Zhao 2004), and in African countries (e.g. Khanchel 2009;

Ongore 2011).

– Level of management. Most studies have focused on top management as the

unit of analysis (e.g. Chang and Wong 2003, 2004; Xiaoyang Li 2007;

Walters 1995; Zhao et al. 2010) although individual studies have investigated

the discretion of middle management (e.g. Caza 2007).

– Firm type. Empirical studies have combined firms to various (partly

overlapping) aggregates, such as listed firms (e.g. Huiyuan Chen 2006;

Zhang and Li 2008b; Zhang and Xie 2008), multinational corporations

(e.g. Heinecke 2011; Thomas and Peyrefitte 1996; Wang et al. 2008), inter-

national joint ventures (e.g. Lopez-Navarro and Camison-Zornoza 2003; Yan

et al. 2010), and state-owned enterprises (e.g. Groves et al. 1994; Xu et al.

2005; Zhang 1997).

– Firm size. Units of analysis in studies have ranged from small firms with a

few hundred employees (e.g. Cheng et al. 2006) to large firms with tens of

thousands of employees (e.g. Adams et al. 2005; Werner and Tosi 1995;

Williamson 1963) or combinations of smaller and larger firms.

– In addition to these examples of differences in the unit of analysis, the

managers investigated in the empirical literature have varied in a number of

other ways, such as in terms of industry and time-related aspects.26

Before exploring differences in the next aspect of the research design, the

above review of differences in the unit of analysis between existing studies is

utilised to derive implications for the unit of analysis of the present study. While

the literature review has shown that there have been individual studies for

middle management and China, compared to the total amount of studies, evi-

dence on the impact of discretion on performance has remained particularly

scarce for both middle management (see Caza 2007, p. 1) and for China (see

Yougen Li and Zhao 2003, p. 6; Zhang and Li 2008a, pp. 37–38). It follows that

given the abundance of middle managers in organisations27 and the importance

of China for domestic Chinese firms and foreign multinationals (Aminpour and

Woetzel 2006, p. 41; Grant 2006, p. 25; Hexter 2006, p. 1; Hoover 2006, p. 92;

26 As noted in Chap. 3, a unit of analysis is also defined in terms of time (Northrop and Arsenault

2007, p. 214).27 As explained in Chap. 3, there are tens of thousands of plant managers in China alone (Guojia

tongji ju [National Bureau of Statistics] 2007, 14–1, 14–2, 14–18). With plant managers being but

one example of middle managers, this translates into an even larger number of middle managers in

organisations worldwide.


Kaufmann et al. 2005, p. 21; McGregor 2005, pp. 2, 272; Pascha 1998, p. 57;

Taube 2008, p. 186; Tian 2007, pp. 7–8), this limited evidence on the impact of

discretion motivates the choice of middle management in China as the unit of

analysis (see Sect. 2.4.4 and Chap. 3). In addition, the differences between

existing studies in the unit of analysis’ firm type and firm size lead to the

formulation of Hypotheses 2 and 3—as discussed below in Sect. 2.2.4.

• As with the unit of analysis, the measure of discretion and the measure ofperformance have also varied strongly across existing studies, confirming

differences in these measures as potential causes for the observed differences

in empirical results between studies. The breadth of different discretion

measures employed is exemplified in Table 2.5, which segments different

measures of discretion according to their implicit assumptions regarding the

dimensions of discretion (which have been discussed in Sect. 2.1.2).28 The threerows in Table 2.5 correspond to studies that assume managerial discretion to be

unidimensional (first row) or multidimensional (second row) or that make no

assumption regarding discretion’s dimensionality (third row):

– Unidimensional. As explained in Sect. 2.1.2, most empirical studies have

assumed thatmanagerial discretion is unidimensional in the sense that discretion

(across multiple areas of the manager’s work) can be measured overall by a

single construct (which then is expected to either increase, not alter or decrease

performance). Among these studies, however, there has been considerable

variation in terms of how the assumed unidimensional construct of discretion

is measured. Table 2.5 shows that one possibility is to measure a manager’s

discretion in a single dimension of his/her work and to take this to represent

discretion overall (e.g. Acemoglu et al. 2007).29Another possibility is to gauge a

manager’s discretion across several dimensions of his/her work and combine

these indicators into a single construct of discretion (e.g. Barnabas and Mekoth

2010;Caza 2007;Chang andWong2003;Cheng et al. 2006;Gammelgaard et al.

2010).30 A third possibility is to gauge a single construct of discretion not by

direct measures of the manager’s discretion but rather by proxy measures that

are expected to be empirically related to discretion overall (e.g. Agarwal et al.

2009; Huiyuan Chen 2006; Khanchel 2009; Yougen Li and Zhao 2004;

Williamson 1963; Zhang and Li 2008b).31

28 Relating these measures of discretion to the dimensions of discretion is an important step in the

present study’s attempt to work towards resolving the discretion puzzle and leads to the formula-

tion of Hypothesis 1.29 Acemoglu et al. (2007) measure autonomy for British firms only in the domain of employment

decisions and autonomy for French firms only in the domain of investment decisions in two out of

their three datasets.30 For instance, Caza (2007, pp. 14–16) combines indicators on the manager’s discretion in

training resources, hiring, firing, and assigning specific tasks into a unidimensional discretion

construct to investigate its impact on performance, although he later finds that discretion covers

multiple dimensions (pp. 26–82).31 For example, Agarwal et al. (2009, p. 2221) proxy managerial discretion by the length of lockup,

notice, and redemption periods for the hedge funds under investigation.


– Multidimensional. By contrast, only a small minority of studies have treated

discretion as multidimensional, assuming that multiple constructs are needed

to represent discretion overall and that single constructs can only meaning-

fully represent a specific dimension of discretion (see Sect. 2.1.2). Table 2.5

reveals alternative approaches in the literature. Some scholars have measured

a manager’s discretion in a single dimension of his/her work, but in contrast

to the studies assuming unidimensionality taken this to represent only a

particular dimension of discretion (e.g. Venaik 1999; Xu et al. 2005).

For instance, Venaik’s (1999) construct aims to represent discretion in mar-

keting decisions. Other scholars have measured discretion in multiple

Table 2.5 Measures of discretion in empirical studies

Assumption on dimensionalitya Examples from empirical studiesb

Unidimensional (single construct used to

represent discretion overall)

• Direct indicator(s) within one dimension of

discretion to represent discretion overall

(e.g. Acemoglu et al. 2007)

• Direct indicators across several dimensions of

discretion combined to represent discretion

overall (e.g. Barnabas and Mekoth 2010;

Caza 2007; Chang and Wong 2003; Cheng

et al. 2006; Gammelgaard et al. 2010)

• Proxy indicator(s) to represent discretion

overall (e.g. Agarwal et al. 2009; Huiyuan

Chen 2006; Khanchel 2009; Yougen Li and

Zhao 2004; Williamson 1963; Zhang and

Li 2008b)

Multidimensional (single construct used to

represent a dimension of discretion, thus

multiple constructs used to represent discretion

overall)

• Direct indicator(s) within one dimension of

discretion to represent one specific dimension

of discretion (e.g. Venaik 1999; Xu et al. 2005)

• Direct indicators across several dimensions of

discretion separated into multiple discretion

constructs (e.g. Groves et al. 1994; Xiaoyang

Li 2007)

• Proxy indicator(s) to represent one specific

dimension of discretion (e.g. Thomas and

Peyrefitte 1996)

No assumption (discretion not mentioned) • E.g. diffusion of ownership, managerial stock

ownership, and internal representation on the

board (e.g. Baysinger and Butler 1985;

Chaganti et al. 1985; Demsetz and Lehn 1985;

Donaldson and Davis 1991; Kesner 1987;

Palmer 1973; Zahra and Stanton 1988)

a Sect 2.1.2 provides a discussion of the dimensions of discretion. Moreover, as explained in

Sect. 2.1.2, the literature draws on studies that have labelled discretion by similar, related terms,

such as autonomyb A direct indicator here stands for a single measure (i.e. indicator) of a manager’s discretion in a

single area (i.e. domain) of his/her work. A proxy indicator here stands for a single measure (i.e.

indicator) that does not measure a manager’s discretion in a single area but rather is expected to be

related to his/her overall discretion


dimensions but unlike the studies assuming unidimensionality differentiated

these indicators into multiple constructs for separate dimensions of discretion

(e.g. Groves et al. 1994; Xiaoyang Li 2007). For example, Xiaoyang Li

(2007) uses separate constructs for investment decision discretion and labour

decision discretion. Finally, scholars employing proxy measures have at

times labelled their constructs as referring to a particular discretion dimension

rather than discretion overall (e.g. Thomas and Peyrefitte 1996).32

– No assumption. There are some studies that do not explicitly aim to estimate the

impact of discretion on performance (and thereforemake no assumption regard-

ing discretion’s dimensionality), but that have, nevertheless, been interpreted as

offering evidence on the discretion-performance relationship (e.g. Chang and

Wong 2003, p. 7). The reasoning is that these studies measure the performance

impact of constructs that may be related to discretion, such as diffusion of

ownership, managerial stock ownership, and internal representation on the

board (e.g. Baysinger and Butler 1985; Chaganti et al. 1985; Demsetz and

Lehn 1985; Donaldson and Davis 1991; Kesner 1987; Palmer 1973;

Zahra and Stanton 1988). These studies are included in Table 2.5 to generate

extensive implications in Sect. 7.3.1.33

While not the focus of the present study, it should be noted that there have

also been differences in the measures of performance across empirical studies on

the impact of discretion on performance. The studies have predominantly

adopted financial measures of firm performance, such as return on assets

(ROA) and to a lesser extent sales growth or Tobin’s Q (e.g. Barnabas and

Mekoth 2010; Bowen et al. 2008; Chang and Wong 2003; Chen 2006; Khanchel

2009; Li 2007; Li and Zhao 2004; Xu et al. 2005; Zhang and Li 2008b). In

addition, some scholars have used latent variables with financial and/or non-

financial indicators as measures of firm or unit performance (e.g. Caza 2007;

2011; Gammelgaard et al. 2010; Venaik 1999). For example, Caza (2007)

measures the performance of research and development units as a latent variable

with four non-financial reflective indicators, namely an evaluator’s rating of the

R&D unit’s innovation, quality, success in reaching R&D goals, and

contributions to the field of science.

32 Thomas and Peyrefitte (1996) investigate the impact of discretion on performance using

Finkelstein and Hambrick’s (1990) industry-level discretion, which has been further advanced in

subsequent studies (e.g. Abrahamson and Hambrick 1997; Datta et al. 2003; Hambrick and

Abrahamson 1995; Hambrick et al. 1993; Keegan 2006; Keegan and Kabanoff 2008).33 The value of the review of existing measures of discretion in the literature (summarised in

Table 2.5) extends beyond demonstrating that the measures of discretion have differed starkly in

the literature. Relating these measures of discretion to the dimensions of discretion is an important

step in the present study’s attempt to work towards resolving the discretion puzzle and leads to the

formulation of Hypothesis 1 in Sect. 2.2.4. Based on the test results of Hypothesis 1 (see Sect.

6.2.1), Sect. 7.3.1 generates implications on how these measures of discretion in Table 2.5 might

be flawed and partly responsible for the discretion puzzle. This leads to recommendations on how

discretion measures should be built and interpreted in future research.


• In addition to the unit of analysis and measures of discretion and performance,

the comparison of research designs between studies confirms that there have

been noteworthy differences in the modelling methodology for estimating the

impact of the measure of discretion on the measure of performance for the

chosen unit of analysis. These aspects of the research design might therefore

also constitute potential causes for the observed differences in empirical results

between existing studies. As to modelling methodology, the extant studies

principally differ in terms of the equations estimated rather than the multivariateanalysis techniques adopted for estimating these equations:34

– Multivariate analysis techniques. The impact of managerial discretion on

performance in existing studies has often been estimated either by means of

multiple regressions, such as ordinary least squares (e.g. Barnabas and

Mekoth 2010; Chang and Wong 2003; Khanchel 2009), or by means of

variance-based and covariance-based structural equation models (e.g. Caza

2011; Gammelgaard et al. 2010; Heinecke 2011; Venaik 1999).

– Equations. Studies have varied broadly as regards the terms included in the

equation for modelling discretion’s impact on performance. In terms of the

notation of Equation (2.1) (see Sect. 2.2.1), studies have often estimated a

single linear direct effect of discretion on performance ( d ) with several

control variables (C ) and without moderator variables (M ) (e.g. Agarwal

et al. 2009; Bowen et al. 2008; Chang and Wong 2004; He et al. 2009; Stano

1976; Werner and Tosi 1995; Zhang 1997). The few studies that have treated

discretion as multidimensional have estimated multiple linear direct effects of

discretion (e.g. d1 and d2 ) (e.g. Groves et al. 1994; Xiaoyang Li 2007).

Furthermore, a few scholars have modelled non-linear direct effects of

discretion (e.g. Huiyuan Chen 2006; Zhang and Li 2008b).35 Finally, a

minority of studies have included moderator variables (M), as explained in

Sect. 2.2.4 (e.g. Caza 2007; Cheng et al. 2006; Yougen Li and Zhao 2004;

Lopez-Navarro and Camison-Zornoza 2003; Wang et al. 2008; Xu et al.

2005; Zhao et al. 2010).

In sum, the exploration of existing empirical evidence has revealed that the

estimated impact of managerial discretion on performance has varied strongly

between studies (ranging from positive to neutral and even negative) and that the

research design of those studies has also varied strongly (e.g. unit of analysis,

measures of discretion and performance, and modelling methodology). However,

as research designs have simultaneously differed in many ways between studies, it

is difficult to infer whether any particular difference in the research design (e.g. the

measure of discretion) could account for the differences in empirical results while

34 Section 2.2.1 provides an explanation of these two aspects of the modelling methodology.35 For instance, Zhang and Li (2008b, p. 122) as well as Chen (2006) include a quadratic term so as

to model non-linear effects, i.e. an inversed U-shaped relationship between discretion and perfor-

mance, with discretion initially increasing and eventually decreasing performance.


holding all other differences in the research design constant.36 A solution is

therefore needed for testing whether a given difference in the research design

might have contributed to the divergent results in the literature on the impact of

discretion on performance that give rise to the discretion puzzle. This solution is

discussed in the next section, i.e. Sect. 2.2.4.

2.2.4 Implications for Hypotheses

The present study offers a solution for testing whether a given difference in the

research design (e.g. the measure of discretion) might have contributed to the

divergent results on the impact of discretion on performance (i.e. the discretion

puzzle). The solution lies in testing whether an individual potential cause

(i.e. a difference in the research design identified above from comparisons betweenempirical studies) produces different empirical results within a single empirical

study so as to hold all the other potential causes (i.e. other aspects of the research

design) constant. This approach makes it possible to test hypotheses on whether

individual differences in research designs between studies can influence the

estimated impact of discretion on performance while holding all other aspects of

the research design constant.

These hypotheses can be tested by building a new empirical discretion model

(see Chap. 4) that exhibits a higher degree of granularity than the models in many

of the aforementioned studies. This granularity refers to the new model

differentiating between different aspects inherent in existing research designs

(e.g. the measure of discretion or the unit of analysis) in a more fine-grained way

than many empirical studies have done before. For example, instead of combining

different indicators of discretion into unidimensional constructs of discretion (e.g.

Barnabas and Mekoth 2010; Caza 2007; see Table 2.5; Chang and Wong 2003;

Cheng et al. 2006; Gammelgaard et al. 2010), the new model differentiates discre-

tion into multidimensional constructs and can therefore test whether different

measures of discretion (i.e. different research designs) can have different impacts

on performance (i.e. different empirical results).

36Most of the studies above differ from each other in terms of both their unit of analysis and their

discretion measure, and in some cases further differ in terms of their performance measure and

modelling methodology. When attempting to hold the unit of analysis constant by comparing only

studies on top management in China, it is found that the contradictory empirical results remain—

with Chang and Wong (2003, 2004) and Zhang (1997) finding a positive, Li and Zhao (2004)

finding a neutral, and Xu et al. (2005) finding a negative direct effect of managerial discretion on

performance. While this hints that controlling for the unit of analysis in terms of level of

management (i.e. top management) and geography (i.e. China) does not (on its own) account for

the literature’s contradictory results, such inferences are hindered by the fact that the studies’

research designs still differ in a number of other ways (e.g. discretion measures and modelling

methodologies).


Likewise, as explained below, the new model differentiates the unit of analysis

by firm type and firm size, which makes it possible to test whether units of analysis

differing in type and size (i.e. different research designs) can have different impacts

on performance. The present study’s approach therefore transfers the differences in

research designs between existing empirical studies into a more granular model

within a new empirical study. This makes it possible to infer whether a particular

identified potential cause (i.e. a difference in research designs) might be responsible

for differences in estimated impacts of discretion on performance—and thereby

contributes towards resolving the discretion puzzle.37

Having outlined the approach taken in this study, the study’s hypotheses are

derived below from the potential causes identified in Sect. 2.2.3 by comparing the

research designs between existing studies. Differences in the unit of analysis, themeasures of discretion and performance, and the modelling methodology were all

identified as potential causes for differences in the estimated impacts of discretion.

These potential causes lead to the formulation of three hypotheses for this study:

• Differences in the measure of discretion across studies and formulation ofHypothesis 1 (dimensions of discretion).38 The above review of existing studies

has shown that various different measures of discretion have been employed in

empirical research to date, as summarised in Table 2.5 above. This empirical

analysis has suggested that differences in the measure of discretion between

studies might potentially help explain the differences in the results between the

studies (ranging from positive to neutral and negative impacts of discretion on

performance). Moreover, from the theoretical analysis in Sect. 2.1.2 (grounded

mostly in managerial discretion theory; see Sect. 2.3.1), there is reason to

believe that managerial discretion consists of several types, i.e. is

37 It should be noted that as explained in Box 1.1 in Sect. 1.2 on the delimitations of the research

objective, the present study’s approach is designed as a proof-by-counter-example and is subject to

the caveat of observational cross-sectional studies in terms of demonstrating causality (e.g.

Caza 2007, p. 46; Finkelstein and Hambrick 1990, p. 500; Granger 1969; Sanchez 2008, p. 5;

Simon 1954, pp. 477–478; Wagner 2002, pp. 287–292; see Sect. 5.4.1). In simple terms, if it is

found in the instance of the present study that discretion is multidimensional, this could constitute

a proof-by-counter-example that discretion is not unidimensional in every case. While this does

not prove that a unidimensional measure of discretion is flawed in the existing studies and

responsible for the discretion puzzle, it would offer evidence that this is a possibility and warrants

further investigation.38 From an empirical point of view, differences in themeasure of performance across studies couldlikewise offer a potential explanation for the divergent empirical results on the impact of discretion

on performance. Some of the existing literature can be interpreted in favour of the hypothesis that

differences in measures of performance may help explain differences in empirical results. For

example, Khanchel (2009) finds a significantly positive impact of discretion on return on assets

(ROA) but insignificant results when measuring performance by Tobin’s Q. Likewise, Wang et al.

(2008) find a positive impact of discretion on sales efficiency but a negative impact on strategic

partnership with the global customers. While this potential explanation lies beyond the scope of

the present study, future studies might extend the present study’s approach of distinguishing

measures in a more granular way and formulate a hypothesis analogous to Hypothesis 1 for testing

whether differences in performance measures help explain the discretion puzzle.


multidimensional (e.g. Barnabas and Mekoth 2010; Carpenter and Golden 1997,

p. 195; Caza 2007, pp. 26–82; Chen 2006; Finkelstein and Peteraf 2007, p. 245;

Groves et al. 1994, p. 190; Hambrick and Abrahamson 1995, p. 1439;

Hambrick and Finkelstein 1987, pp. 371–402; Hambrick et al. 1993, p. 409;

March and Simon 1958; Perrone et al. 2003, pp. 422–423). Combining these

theoretical and empirical points of view, it is conceivable that discretion is

multidimensional in that discretion granted in a certain domain of the manager’s

work (e.g. making capital investments) improves performance and discretion in

another area (e.g. hiring workers) reduces performance—and further, that a

reason why empirical studies with different discretion measures have produced

different results might be that these measures have tapped into distinct

dimensions of discretion. This proposition can be verified by testing whether

discretion is unidimensional (i.e. null hypothesis H01 ) versus multidimensional

(i.e. alternative hypothesis H11 ) in its impact on performance (i.e. whether the

impact of discretion on performance differs between dimensions of discretion):




The present study’s first hypothesis (as already derived in Sect. 2.1.2) is

thus further substantiated by the existing empirical evidence on the impact

of discretion on performance, which shows that measures of discretion have

differed between existing studies and that few existing studies have adopted the

greater granularity to allow for a multidimensional construct of discretion.39 If

H01 (i.e. unidimensionality) can be rejected in favour of H1

1 (i.e. multi-

dimensionality), this would provide significant evidence that the performance

impact of discretion can depend on the area of work in which a manager is

granted discretion. This would help resolve the discretion puzzle by showing

that existing studies might have estimated positive, neutral, and negative

impacts of discretion on performance because their measures of discretion

have tended not to distinguish properly between the dimensions of discretion,

i.e. the areas of the manager’s work in which discretion is granted.

• Differences in the unit of analysis and modelling methodology across studiesand formulation of Hypothesis 2 (firm type) and Hypothesis 3 (firm size).

39 As explained in Sect. 2.1.2, most existing studies have adopted unidimensional measures of

discretion, meaning they have not exhibited sufficient granularity to allow for a multidimensional

measure of discretion (e.g. Agarwal et al. 2009; Barnabas and Mekoth 2010; Berman et al. 2005;

Bloom et al. 2009a, b; Cameron 2000; Caza 2007; 2011; Chang and Wong 2003; Chen 2006;

Cheng et al. 2006; Finkelstein and Boyd 1998; Finkelstein and Hambrick 1990; Gammelgaard

et al. 2010; He et al. 2009; Khanchel 2009; Li and Zhao 2004; Magnan and St-Onge 1997;

Marin and Verdier 2006; Rajagopalan and Finkelstein 1992; Williamson 1963; Zhang and Li

2008b; Zhang et al. 2006a, b).


The empirical analysis above has demonstrated that in addition to the measure of

discretion, the unit of analysis and modelling methodology differ between

existing studies and thus might help explain the observed differences in empiri-

cal results between studies (see Sect. 2.2.3). As to the unit of analysis, studies

with distinct results have analysed distinct managers, with the unit of analysis

differing in such ways as the level of management (e.g. Caza 2007; Chang and

Wong 2003; 2004; Xiaoyang Li 2007; Walters 1995; Zhao et al. 2010), geogra-phy (e.g. Agarwal et al. 2009; Barnabas and Mekoth 2010; Caza 2011;

Crossland and Hambrick 2007; Gammelgaard et al. 2010; He et al. 2009;

Khanchel 2009; Yougen Li and Zhao 2004; Ongore 2011; Stano 1976;

Tang 2008; Venaik 1999), firm type (e.g. Chen 2006; Groves et al. 1994;

Heinecke 2011; Lopez-Navarro and Camison-Zornoza 2003; Thomas and

Peyrefitte 1996; Wang et al. 2008; Xu et al. 2005; Yan et al. 2010; Zhang and

Li 2008b; Zhang and Xie 2008; Zhang 1997), and firm size (e.g. Adams et al.

2005; Cheng et al. 2006; Werner and Tosi 1995; Williamson 1963). It follows

that one way to analyse the effect of differences in the unit of analysis is to test

whether differences in firm types and/or firm sizes (or any of the other influences

on managers) affect (i.e. moderate) the impact of discretion on performance

while holding the other aspects of the unit of analysis constant (i.e. fixing the

level of management at middle management and fixing the geography at China).

In addition to this empirical motivation, there is theoretical rationale for testing

whether influences on the manager (e.g. firm type and firm size) moderate the

impact of discretion on performance: As explained in Box 2.6 in Sect. 2.4.2, the

empirical and theoretical literature combined motivate this study to test

the hypotheses that the influences on managers of firm type and firm size affect

the way in which managers use discretion and thus the impact of discretion:

Hypothesis 2 (Firm Type)

H02: The impact of managerial discretion on performance is equal between different

firm types (i.e. domestic Chinese firms versus foreign multinationals).

H12: The impact of managerial discretion on performance differs between different


Hypothesis 3 (Firm Size)

H03: The impact of managerial discretion on performance is equal between different

firm sizes.


firm sizes.

If the null hypotheses (H02 and H0

3 ) can be rejected in favour of the alternative

hypotheses (H12 and H1

3), there would be significant evidence that the performance


impact of discretion can depend on the influences on the manager in terms of firm

type and firm size. This would contribute to the resolution of the discretion puzzle by

indicating that existing studies might have estimated positive, neutral, and even

negative impacts of discretion on performance because their units of analysis and

modelling methodologies have often failed to differentiate sufficiently the impact

of discretion between different influences on the managers, such as firm type and

firm size.

It should be noted that testing Hypotheses 2 and 3 (which were derived from

the unit of analysis as a potential cause of differences in empirical results) also

tests for differences in the modelling methodology as a potential cause of

differences in empirical results—i.e. differences in the moderator variables (M)included in Equation (2.1) in Sect. 2.2.1 when estimating the impact of manage-

rial discretion (D) on performance (P).40 The equivalence of the two explanationscan be understood by considering the example of firm type from Hypothesis 2. It

was explained that one may test for whether differences in the unit of analysis in

terms of firm type can affect the estimated impact of discretion on performance

(i.e. empirical results) by testing whether or not the impact of discretion on

performance differs between different firm types (or more formally, testing H02

against H12). This is equivalent to testing whether or not firm type moderates the

impact of discretion on performance (i.e. has a moderating effect) and thus

whether including firm type as a moderator variable (M) can potentially change

the estimated impact of discretion on performance (i.e. empirical results).41 The

argument extends to differences in the influences on the manager in terms of firm

size. Testing Hypothesis 3 for whether differences in the unit of analysis in terms

of firm size affect the impact of discretion on performance (i.e. empirical results)

simply tests for whether firm size moderates the impact of discretion on

performance.

40 As explained above, the reviewed empirical studies have also differed from each other in terms

of aspects of the modelling methodology other than moderator variables, such as multivariate

analysis techniques. Future studies could investigate these differences as potential explanations of

the discretion puzzle.41 This equivalence can be understood by drawing on Chap. 4. With the multi-group comparison

approach for testing moderating effects (see Sect. 4.3.1; e.g. Arnold 1982; Henseler and Fassott

2010, pp. 719–721; Rigdon et al. 1998, p. 1; Venkatraman 1989, p. 426), Hypothesis 2 is tested by

separately estimating an equation similar to Equation (2.1) in Sect. 2.2.1 for the two firm type

groups, which yields a separate direct effect of discretion on performance for the unit of analysis of

Chinese firms (dChinese) and a separate one for multinationals (dMulti:). Testing whether firm type

affects the impact of discretion on performance then tests whether the difference between dChinese

anddMulti: is significant, and this difference is the moderating effect of firm type (see Sect. 4.3.2). In

principle, the product term approach for testing moderating effects (see Sect. 4.3.1; e.g. Carte and

Russell 2003, pp. 480–495; Chin et al. 2003, pp. 196–200; Irwin and McClelland 2001, p. 105) can

also be applied for testing Hypothesis 2, whereby firm type is coded as a dichotomous categorical

variable by using a dummy moderator (Henseler and Fassott 2010, p. 721; Qureshi and Compeau

2009, p. 199). With reference to Equation (2.1) in Sect. 2.2.1, M ¼ 0 could represent multina-

tionals and M ¼ 1 Chinese firms, so the moderating effect m would measure the difference in the

impact of discretion on performance between a Chinese firm and a multinational, and testing for

m 6¼ 0 would test Hypothesis 2.


The insight that moderator variables (i.e. modelling methodology) differentiate

the impact of discretion on performance with respect to an aspect of the unit of

analysis (such as firm type or firm size) can be used to interpret existing empirical

evidence that has used moderator variables (M) in their variants of Equation (2.1)

in Sect. 2.2.1. As Equation (2.2) in Sect. 2.2.1 has shown, in studies with

moderators the total impact of discretion on performance is the sum of the direct

effect of discretion (d ) and the moderating effect multiplied by the value of the

moderator variable (m �M). Viewing the existing empirical literature in this light

then further strengthens the present study’s motivation to test Hypotheses 2 and 3

with greater granularity:

• The fact that existing studies have rarely modelled moderator variables (M )

means that they have seldom differentiated the impact of discretion on perfor-

mance by different aspects of the unit of analysis (such as firm type or firm size;

e.g. Agarwal et al. 2009; Barnabas and Mekoth 2010; Bowen et al. 2008; Chang

and Wong 2003; Gammelgaard et al. 2010; Groves et al. 1994; He et al. 2009;

Khanchel 2009; Li 2007; Stano 1976; Venaik 1999; Werner and Tosi 1995;

Zhang 1997). In conjunction with the finding that existing studies have mostly

adopted unidimensional measures of discretion (see Sect. 2.1.2), this means that

empirical research to date has tested models with limited granularity, rarelydifferentiating the impact of discretion on performance by dimensions of discre-

tion (i.e. measures of discretion) or by influences on managers (i.e. units of

analysis and moderator variables).

• The few studies that have investigated moderator variables have provided

evidence that the impact of discretion on performance can differ according to

different moderators of the unit of analysis (e.g. Caza 2007; 2011; Cheng et al.

2006; Lopez-Navarro and Camison-Zornoza 2003; Wang et al. 2008; Zhang and

Li 2008b; Zhao et al. 2010). This reinforces the rationale for Hypotheses 2 and 3that discretion’s impact on performance may differ by at least some moderators

of the unit of analysis.

• Finally, as the moderator variables investigated to date have focused on aspects

of the unit of analysis other than firm type and firm size, the selection of firm typeand firm size as the influences on managers for Hypotheses 2 and 3 is vindicated.The reason is that in this way, the present study contributes to the literature on

discretion evidence about rarely tested moderators (i.e. firm type and firm size)

rather than rehashing existing moderators.42

42 Based on the discussion in this section, the existing empirical evidence on moderating effects (m)can now be interpreted as showing that the way in which a manager uses his/her discretion can

depend on various influences on managers—in terms of such statistically significant moderators as

managerial experience (Caza 2007), managerial incentives (Cheng et al. 2006), top managementpay gap (Zhang and Li 2008b), corporate control via performance monitoring, incentive systems,

and social integration (Wang et al. 2008), export joint venture group composition (Lopez-Navarroand Camison-Zornoza 2003), and market competition (Zhao et al. 2010). For some of the

moderators tested in the literature, moderating effects were found to be insignificant, such as

managerial education, managerial commitment, and the number of similar units (Caza 2011),

ownership concentration (Yougen Li and Zhao 2004), and the organisational type of state-owned

enterprises in China (Xu et al. 2005), i.e. employee-owned stock cooperatives, limited liability

companies, and limited liability stock companies.


In sum, this section has conducted a thorough, in-depth review of empirical studies

on the impact of managerial discretion on performance, which completes further cells

in the literature review summary Table 2.6. It explained how empirical studies consist

of a research design (i.e. unit of analysis, measures of discretion and performance, and

modelling methodology) as well as empirical results (i.e. a total impact of discretion on

performance composed of direct and moderating effects). The empirical results of the

existing studies were then found to be contradictory in that discretion had been

estimated to have positive, neutral, and even negative impacts on performance. This

gives rise to the discretion puzzle, which the present study seeks to address. In order to

Table 2.6 Literature review summary (Sect. 2.2 completed)

Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality







Impact on

performance:

• Direct

• Moderating

Contradictory

evidence

• Positive,

neutral,

negative

• Moderators

potentially

important

but rarely

modelled

(esp. firm

type and

firm size)

Unit of analysis Rarely middle

management,

rarely China

Section 2.3

Section 2.4

Hypothesis 1

Hypothesis 2

Hypothesis 3

Middle

management

in China


explore potential causes for the contradictory results, the research designs of the

existing studies were compared, which confirmed that differences between studies in

the unit of analysis, the measures of discretion and performance, and the modelling

methodology could all be potential causes of the observed differences in results. In

order to test whether these potential causes can help explain the discretion puzzle, three

hypotheses were formulated for testing the causes derived from the differences

between existing studies within the present study. These hypotheses refer to the

dimensions of discretion (Hypothesis 1), firm type (Hypothesis 2), and firm size

(Hypothesis 3), and are entered into the literature review summary Table 2.6. Each

hypothesis has empirical as well as theoretical origins, and they collectively postulate

that managers may use their managerial discretion differently depending on the area of

their work in which discretion is granted (i.e. the dimension of discretion) and the

influences onmanagers that are in place (e.g. firm type and firm size). Finally, Table 2.6

notes the selection of the study’s unit of analysis that results from the above discussion.

2.3 Impact of Managerial Discretion on Performancein Existing Theories

This section conducts a thorough review of the theoretical literature on the impact

of managerial discretion on performance, i.e. managerial discretion theory (see

Sect. 2.3.1), principal-agent theory (see Sect. 2.3.2), and stewardship theory (see

Sect. 2.3.3). Each of the three theories is first described briefly in terms of its

historical context, unit of analysis, assumptions, and predictions before it is

reviewed in detail with respect to analysing the impact of discretion on perfor-

mance.43 The theoretical analysis of the impact of discretion on performance covers

the construct of discretion (i.e. definition and dimensionality), the antecedents of

discretion, and the consequences of discretion (i.e. direct effects and moderating

effects), which are then summarised in the literature review summary Table 2.7.

In this way, the review works towards fulfilling the four purposes laid out in the

opening section of Chap. 2:

1. As to the research gap and research objective (see Sects. 1.1 and 1.2), the reviewdemonstrates that the existing theories do not unequivocally explain the contra-

dictory empirical results on the impact of discretion on performance from the

previous section, which gives rise to the discretion puzzle.

2. As to the postulate and hypotheses (see Sect. 2.4), the review of the three

theories culminates in formulating the study’s final hypothesis (Hypothesis 4),

and reinforces Hypothesis 1, Hypothesis 2, and Hypothesis 3, derived above.

43 The theories can be applied to study a wide variety of topics other than the impact of managerial

discretion on performance, such as a range of topics in economics and finance for principal-agent

theory. However, as these applications do not contribute to fulfilling the present study’s research

objective, they are not included in this literature review. Instead, the theories are reviewed only

with respect to what is relevant to this study.

2.3 Impact of Managerial Discretion on Performance in Existing Theories 57

3. As to the unit of analysis (see Chap. 3), this section verifies that the theories are

applicable to the study’s unit of analysis and utilises this unit for examples.

4. As to the study’s model (see Chap. 4), this section distils the theoretical content

of the relevant theories for subsequent integration into the present study’s new

model on the impact of managerial discretion on performance.

These results enter into the grey-shaded cells in the literature review summary

Table 2.7 and thereby complete the summary of the empirical and theoretical

literature. This allows Sect. 2.4 to synthesise the current state of research, i.e. the

composite findings of the relevant literature on the impact of discretion on perfor-

mance across the columns in Table 2.7 into the research gap and research objective,


Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality







Impact on

performance:

• Direct

• Moderating

Contradictory

evidence

• Positive,

neutral,

negative

• Moderators

potentially

important

but rarely

modelled

(esp. firm

type and

firm size)


management,

rarely China

Section 2.3

Section 2.4

Hypothesis 1

Hypothesis 3

Middle

management

in China

Hypothesis 2


the postulate and hypotheses, and the unit of analysis. Then, the completed review

will allow Chap. 4 to develop a new discretion model in a way that builds on the

available empirical and theoretical literature and permits testing the study’s

hypotheses for its unit of analysis so as to fulfil its research objective.

Before delving into the details of the theoretical literature, each of the three

theories reviewed below is outlined briefly for the reader in terms of how it

contributes to explaining the impact of managerial discretion on performance.

While managerial discretion theory contributes strongly to specifying the theoreti-

cal context for investigating the impact of discretion, principal-agent theory and

stewardship theory postulate mechanisms that explain why the impact of discretion

on performance could be positive or negative and in which way it might be

moderated:

• Managerial discretion theory44 (see Sect. 2.3.1) specifies the theoretical contextof managerial discretion by defining the construct of discretion, postulating its

antecedents (i.e. its causes, determinants or sources), and postulating its

consequences (i.e. what outcomes discretion may affect). While it thereby

specifies that discretion can have an impact on performance, the nature of this

impact (e.g. positive versus negative) remains unspecified (e.g. Abrahamson and

Hambrick 1997; Berman et al. 2005; Carpenter and Golden 1997; Caza 2007;

2011; Huiyuan Chen 2006; Crossland 2007; Crossland and Hambrick 2007;

Datta et al. 2003; Finkelstein and Boyd 1998; Finkelstein and Hambrick 1990;

Finkelstein and Peteraf 2007; Haleblian and Finkelstein 1993; Hambrick and

Abrahamson 1995; Hambrick and Finkelstein 1987; Hambrick et al. 1993;

Hutzschenreuter and Kleindienst 2007; Keegan 2006; Keegan and Kabanoff

2008; Key 2002; Yougen Li and Zhao 2004; Magnan and St-Onge 1997;

Quigley and Hambrick 2009; Rajagopalan and Finkelstein 1992; Thomas and

Peyrefitte 1996; Zhang and Li 2008b; Zhang et al. 2006a, b).

• Principal-agent theory (see Sect. 2.3.2), by contrast, does not fully specify the

theoretical context of managerial discretion in terms of defining discretion, its

antecedents, and its consequences. Instead, the theory postulates that managers

(agents) tend to use discretion to pursue their own interests at the cost of the

principal and thus in most cases to the detriment of performance (e.g. Agrawal

and Knoeber 1996; Baysinger and Butler 1985; Berger et al. 1997; Brush et al.

2000; Chang and Wong 2003; Childs and Mauer 2008; Denis et al. 1997;

Eisenhardt 1989; Fama 1980; Fama and Jensen 1983a, b; He et al. 2009;

Jensen 1986; Jensen and Meckling 1976; Jensen and Murphy 1990; Jensen and

Ruback 1983; Laffont and Martimort 2002; Lang et al. 1995; Levinthal 1988;

44 As described in Sect. 2.3.1, managerial discretion theory was originally developed as a reconcili-

ation of population ecology (e.g. Aldrich 1979; Baum 1996; Baum and Amburgey 2002; Carroll

1988; Freeman et al. 1983; Hannan and Freeman 1977, 1984; Singh and Lumsden 1990; Tushman

and Romanelli 1985; Zohar and Luria 2005) and strategic choice theory (e.g. Child 1972, 1997,

2002; Child et al. 2003; Elbanna and Child 2007; Hitt and Tyler 1991; Hrebiniak and Joyce 1985;

Judge and Zeithaml 1992; Marlin et al. 1994; Miles and Snow 1978; Stienstra et al. 2004).


Ongore 2011; Shleifer and Vishny 1997; Spremann 1987; Thepot 2007;

Thomsen and Pedersen 2000; Walters 1995; Wang et al. 2008; Weidenbaum

and Jensen 1993; Werner and Tosi 1995, p. 1673; Xu et al. 2005; Zou 1989).

Principal-agent theory therefore mostly implies a negative impact of managerial

discretion on performance.45

• Stewardship theory (see Sect. 2.3.3) likewise does not focus on specifying

discretion’s theoretical context. Instead, it contends that managers (stewards)

‘are good stewards of the corporations and diligently work to attain high levels

of corporate profit and shareholder returns’ (e.g. Albanese et al. 1997;

Arthurs and Busenitz 2003; Corbetta and Salvato 2004; Davis et al. 1997a, b;

Dicke and Ott 2002; Donaldson 1990; Donaldson and Davis 1989, 1991, 1993,

1994, p. 159; Eddleston and Kellermanns 2007; Fox and Hamilton 1994;

Lane et al. 1999; Liu and Cai 2004; Miller and Le Breton-Miller 2006;

Mills and Keast 2009; Muth and Donaldson 1998; Salvato 2002; Tian and Lau

2001; Tosi et al. 2003; Van Slyke 2007; Vargas Sanchez 2001, 2004, 2005;

Zahra 2003). In contrast to principal-agent theory, stewardship theory therefore

specifies a positive impact of managerial discretion on performance.

2.3.1 Managerial Discretion Theory

Following the structure introduced above, the study’s literature review on manage-

rial discretion theory in the present section is conducted in two parts:

• Section 2.3.1.1 describes managerial discretion theory briefly in terms of its

historical context, assumptions, predictions, and unit of analysis.

• Section 2.3.1.2 reviews managerial discretion theory in detail with respect to

analysing the impact of discretion on performance—concerning the construct of

discretion (i.e. definition and dimensionality), the antecedents of discretion, and

the consequences of discretion (i.e. direct effects and moderating effects).

Subsequently, Sects. 2.3.2 and 2.3.3 apply a parallel structure so as to investigate

principal-agent theory and stewardship theory in a comparable manner.

2.3.1.1 Description of TheoryBeginning with the theory’s historical context, managerial discretion theory

(e.g. Abrahamson and Hambrick 1997; Berman et al. 2005; Carpenter and Golden

1997; Caza 2007; 2011; Chen 2006; Crossland 2007; Crossland and Hambrick

45As explained in Sect. 2.3.2, it is possible to assume that the principal is less performance-

maximising than the agent, in which case managerial discretion may positively affect performance

(Chang and Wong 2003, pp. 1–7). However, this assumption is rarely made in principal-agent

theory (Thomsen and Pedersen 2000, p. 690). The literature on principal-agent theory therefore

generally contends that managerial discretion has a negative direct effect on performance

(e.g. Caza 2007, p. 10; Caza 2011; Chang and Wong 2003, p. 7; Davis et al. 1997b, p. 38;

Hutzschenreuter and Kleindienst 2007, p. 4; Jensen and Murphy 1990; Zhao et al. 2010).


2007; Datta et al. 2003; Finkelstein and Boyd 1998; Finkelstein and Hambrick

1990; Finkelstein and Peteraf 2007; Haleblian and Finkelstein 1993; Hambrick and

Abrahamson 1995; Hambrick and Finkelstein 1987; Hambrick et al. 1993;

Hutzschenreuter and Kleindienst 2007; Keegan 2006; Keegan and Kabanoff

2008; Key 2002; Li and Zhao 2004; Magnan and St-Onge 1997; Quigley and

Hambrick 2009; Rajagopalan and Finkelstein 1992; Thomas and Peyrefitte 1996;

Zhang and Li 2008b; Zhang et al. 2006a, b) was formally established by Hambrick

and Finkelstein (1987), who introduced the construct of managerial discretion in

order to reconcile the polar views of organisational outcomes by populationecology and strategic choice theory. Prior to their seminal work, managerial

discretion had been implicitly assumed to be important in determining actions

and organisational outcomes (e.g. Berle and Means 1932; Chandler 1962) and at

times had been explicitly mentioned (e.g. Montanari 1978; Williamson 1963)—yet

it was Hambrick and Finkelstein’s (1987) seminal work that provided an explicit

definition of managerial discretion (see Sect. 2.1.1). The theories that managerial

discretion theory attempts to reconcile—namely population ecology and strategic

choice theory—offer opposing views as to the extent to which organisations (and in

particular managers in the organisations) have control over their destinies:

• Population ecology or organisational ecology (e.g. Aldrich 1979; Baum 1996;

Baum and Amburgey 2002; Carroll 1988; Freeman et al. 1983; Hannan and

Freeman 1977, 1984; Singh and Lumsden 1990; Tushman and Romanelli 1985;

Zohar and Luria 2005) is a branch of evolutionary organisational theory that

assumes that the actions of individuals in organisations have very limited effects

on organisational outcomes or that these actions themselves are effectively

determined by structural forces, so that the survival of organisations is largelydetermined by natural selection rather than adaptation by decision-makers.46

Consequently, organisations are regarded as entities within populations and in

the process of evolution, those organisations with favourable organisational

competences—such as routines of large organisations (e.g. Baum 1996;

Singh and Lumsden 1990) and the reputation and reproducibility of more mature

organisations (e.g. Freeman et al. 1983; Hannan and Freeman 1984;

Stinchcombe 1965)—successfully adjust to environmental structural changes

(e.g. innovations) and survive. Empirical evidence has partly supported and

partly countered some of the initial postulates, for example, that failure rates

decline with organisational age, and this has led to various refinements of the

theory (see Baum 1996, pp. 73–77; Baum and Amburgey 2002, pp. 305–309).

Nevertheless, population ecology continues to emphasise that performance islargely determined by natural selection due to environmental forces andorganisational competences rather than managerial action.

46 For instance, Hannan and Freeman (1977, p. 957) postulate that ‘there are very strong inertial

pressures on structure arising both from internal arrangements (e.g., internal politics) and the

environment (e.g., public legitimation of organizational activity).’


• Strategic choice theory (e.g. Child 1972, 1997, 2002; Child et al. 2003;

Elbanna and Child 2007; Hitt and Tyler 1991; Hrebiniak and Joyce 1985;

Judge and Zeithaml 1992; Marlin et al. 1994; Miles and Snow 1978;

Stienstra et al. 2004) is an extension of contingency theory that reverses the

focus of population ecology by emphasising the roles of managers in shapingconditions and processes that determine an organisation’s fate. The theory

assumes that there is a group of influential decision-makers (i.e. the ‘dominant

coalition’) who actively make strategic choices that determine organisational

outcomes (Datta et al. 2003, p. 102; Pegels et al. 2000, p. 911). In the decision-

making process, these decision-makers engage in an evaluation of the current

situation, which is influenced by e.g. environmental conditions, rewards

expected by resource providers, and the prior ideology of the decision-makers

(Child 1972, p. 18). In strategic choice theory, performance is therefore largelydetermined by managerial action, which in turn is influenced by environmental,organisational, and managerial contingencies.47

In an attempt to reconcile these opposing views of the organisational outcomes of

population ecology and strategic choice theory, Hambrick and Finkelstein (1987,

p. 371) develop and explore the concept of managerial discretion for chief

executives, which has been defined in Sect. 2.1.1. The authors postulate that the

managerial discretion of chief executives varies ‘from very little to a great deal’ and

that the particular extent of discretion is determined by antecedents or ‘factors in the

environment, the organization, and the chief executive’s own attributes’ (1987,

p. 369), as is discussed below. Depending on whether managerial discretion is high

or low, differing organisational outcomes (including performance) are postulated,

and this reconciles the predictions of population ecology and strategic choice theory:

• When chief executives possess low discretion (e.g. risk-averse CEOs in heavily

regulated industries where pricing and other decisions are fixed externally), they

will be constrained in the extent to which they can engage in managerial action

(i.e. strategic choices) that can influence organisational outcomes. Therefore,

when the latitude of managerial action is narrow, performance is largely deter-mined by environmental circumstances rather than managerial action. In the

case of low managerial discretion, Hambrick and Finkelstein (1987, p. 391) thus

47 These contingencies affecting strategic choices and thus organisational outcomes are investigated

in upper-echelon theory. The theory postulates that strategic choices affecting organisational

outcomes are reflections of the cognitive bases and values (i.e. knowledge, ordering, and preferences)

of influential top managers, since under the assumption of bounded rationality (Simon 1957b) the

managers’ cognitive bases and values are pivotal for processing the complex information from

environmental and organisational stimuli (Aragon-Correa et al. 2004; Galavan 2005; Galavan et al.

2009; Hambrick et al. 1993; Hambrick and Mason 1984; Hutzschenreuter and Kleindienst 2007;

Manner 2010; Pegels et al. 2000). It is argued that these cognitive bases and values can bemeasuredby

observablemanagerial background characteristics (e.g. age, education, and socioeconomic roots), and

indeed it has been confirmed that many of these demographics are empirically related to strategic

choices and performance (see Carpenter et al. 2004; Hambrick 2007).


expect a stable strategy and organisational form over time, as well as relatively

stable performance similar to that of other firms in the environment. Hence, lowdiscretion leads to the predictions of population ecology.

• When chief executives possess high discretion (e.g. risk-prone CEOs with a highdegree of variable compensation in an unregulated industry), they will be able to

engage in managerial action (i.e. strategic choices) that impacts on

organisational outcomes. Then, performance is to a large extent determined bymanagerial action. In the case of high managerial discretion, managerial discre-

tion theory (1987, p. 395) thus posits shifting strategies and organisational

forms, with performance fluctuating above or below average levels. Hence,high discretion leads to the predictions of strategic choice theory.This reconciliation of the predictions of population ecology and strategic choice

theory based on Hambrick and Finkelstein’s managerial discretion theory (1987) is

depicted in the fictitious scatter plot in Fig. 2.2 (which resembles Fig. 2.1 from

Sect. 2.2.2) with respect to performance, since the impact of managerial discretion

on performance is the focus of the present study. As described in Sect. 2.3.1.2, the

impact of discretion can likewise be analysed for a number of other consequences,

such as managerial power (Carpenter and Golden 1997), managerial compensation

(Finkelstein and Boyd 1998;Magnan and St-Onge 1997; Rajagopalan and Finkelstein

1992; Werner and Tosi 1995; Wright and Kroll 2002; Zhang and Xie 2008), workers’

incentives (Groves et al. 1994), a successor chief executive officer’s age (Wang

2009), top management team tenure, trust (Perrone et al. 2003), strategic attention

(Abrahamson and Hambrick 1997), environmental commitment (Aragon-Correa

et al. 2004), pricing (Cameron 2000), organisational knowledge creation (Oh 2002),

and research and development (Zhang et al. 2006a, b). In the scatter plot, managerial

discretion is measured on the horizontal axis and the organisation’s performance on

the vertical axis. When managerial discretion is low (left-hand side), managers have

little latitude of action and therefore limited influence (both positive and negative)

on performance. Performance is then mainly determined by environmental

circumstances, as posited by population ecology. Therefore, on the left of Fig. 2.2,

the performance in firms with lowmanagerial discretion is expected to be closer to the

average performance level of firms (which is determined by environmental

circumstances), for there are only weak idiosyncratic influences of individual

managers on performance. Likewise, on the right of Fig. 2.2, the performance in

firms with high discretion is expected to be either above or below average levels due

to the positive or negative influences of managerial action (i.e. strategic choices) on

performance, i.e. strategic choice theory applies. Managerial discretion theory

thereby successfully reconciles the polar views of organisational outcomes of popu-

lation ecology and strategic choice theory (Crossland 2007; Crossland and Hambrick

2007; Finkelstein and Boyd 1998; Hambrick and Finkelstein 1987; Quigley and

Hambrick 2009). As explained in Sect. 4.1.1, this reconciliation is the starting point

for developing the present study’s new discretion model on the impact of managerial

discretion on performance.

While Hambrick and Finkelstein’s (1987) initial seminal contribution to mana-

gerial discretion theory was focused on chief executives as the unit of analysis,


scholars have subsequently extended managerial discretion theory to a broader set

of units of analysis, such as other managers in top management and even middle

management.48 The discretion of these differing units of analysis has been

investigated in both theoretical and empirical research, as exemplified by the

following studies:

• Studies have mostly analysed the managerial discretion of top management(e.g. chief executive officers) relative to external stakeholders, such as

shareholders, in corporate governance (e.g. Abrahamson and Hambrick 1997;

Adams et al. 2005; Aragon-Correa et al. 2004; Carpenter and Golden 1997;

Cennamo et al. 2009; Crossland and Hambrick 2007; Finkelstein and Boyd

1998; Finkelstein and Hambrick 1990; Finkelstein and Peteraf 2007; Haleblian

and Finkelstein 1993; Hambrick and Abrahamson 1995; Khanchel 2009;

Yougen Li and Zhao 2004; Magnan and St-Onge 1997; Rajagopalan and

Finkelstein 1992; Walters 1995; Wang 2009; Werner and Tosi 1995; Wright

and Kroll 2002; Zhang et al. 2006a, b; Zhang and Xie 2008; Zhao et al. 2010).

High Performance

LowManagerialDiscretion

HighManagerialDiscretion

Low Performance

Fictitious performance-discretion data points for population ecology

Fictitious performance-discretion data points for strategic choice theory

Prediction ofPopulation Ecology

Prediction ofStrategic Choice Theory

Fig. 2.2 Fictitious scatter plot for managerial discretion theory

Source: The author’s own synthesis of managerial discretion theory, population ecology, and

strategic choice theory

48 Outside the scope of managerial discretion theory, notions similar to managerial discretion have

been applied to non-managerial units of analysis. For example, the job design literature (e.g.

Hackman et al. 1975) is concerned with the autonomy of workers from a motivational perspective

(i.e. their discretion relative to management). Moreover, the Nobel Laureates Kydland and

Prescott (1977) have derived far-reaching implications for the design of macroeconomic policy

from their analysis of the discretion of policymakers.


• A minority of studies have explored the discretion of middle managementrelative to top management—either explicitly terming this ‘managerial discre-

tion’ or using related concepts such as ‘autonomy’ (e.g. Acemoglu et al. 2007;

Barnabas and Mekoth 2010; Bloom et al. 2008; Caza 2007; 2011; Glaister et al.

2003; Lopez-Navarro and Camison-Zornoza 2003; Oh 2002; Perrone et al. 2003;

Venaik 1999; see Sect. 2.1.1).

These studies confirm that managerial discretion theory can be applied to various

hierarchical levels in organisations, including the middle management level that is

analysed in the present study (see Caza 2007, pp. 7–8). In order to explain how the

theory can be applied to the middle management analysed in this study (i.e. the

plant manager in China; see Chap. 3), the exploration of the impact of discretion on

performance in Sect. 2.3.1.2 chooses this unit of analysis for giving examples.

2.3.1.2 Analysis of Impact of Discretion on PerformanceHaving outlined managerial discretion theory in general terms (e.g. historical

context, assumptions, predictions, and unit of analysis), this section turns to

reviewing the theory with respect to its explanation of the impact of discretion on

performance. The review shows that, on the one hand, managerial discretion theory

contributes strongly to specifying the theoretical context for investigating the

impact of discretion on performance, i.e. by defining the construct of discretion(see Sect. 2.1), postulating its antecedents (i.e. its causes, determinants or sources),

and generally postulating its consequences (i.e. what outcomes discretion may

affect). This theoretical context is pivotal for building the present study’s new

discretion model in Sect. 4.1.1 and for formulating Hypothesis 1 (dimensions of

discretion) in Sect. 2.1.2. On the other hand, however, while managerial discretion

theory thereby specifies that discretion can have an impact on performance, it does

not specify the nature of this impact, i.e. it does not postulate mechanisms that

would explain why the impact of discretion on performance could be either

positive, neutral or negative and in which way this impact may be moderated.

These issues on the impact of discretion on performance in managerial discretion

theory are discussed below in the following sequence:

1. Construct of managerial discretion (i.e. definition and dimensionality).

2. Antecedents of managerial discretion (i.e. its causes, determinants or sources).

3. Consequences of managerial discretion (i.e. what outcomes discretion may

affect, and in particular the nature of the impact of discretion on performance).

First, as to the construct of discretion, managerial discretion theory makes a

major contribution to the literature in terms of defining discretion and also comes

closest among the three theories reviewed herein to specifying discretion’s

dimensionality (see Sect. 2.1). As to the definition, managerial discretion theory

defines managerial discretion as the ‘latitude of managerial action’, i.e. the extent to

which a manager has multiple courses of action across various domains of his/her

work that he/she is aware of and that are acceptable to the parties that possess power

to constrain the manager (Hambrick and Finkelstein 1987, pp. 371–378; see

Sect. 2.1.1). This definition has been widely accepted in the literature (e.g.

Abrahamson and Hambrick 1997, p. 513; Carpenter and Golden 1997, p. 187;

Caza 2007, p. 27; Chang and Wong 2003, p. 2; Crossland 2007, p. 1; Crossland


and Hambrick 2007, p. 767; Finkelstein and Boyd 1998, p. 179; Finkelstein and

Hambrick 1990, p. 484; Hambrick and Abrahamson 1995, p. 1427; Rajagopalan

and Finkelstein 1992, p. 32). For the unit of analysis of the present study (defined in

Chap. 3), managerial discretion measures the extent to which the plant manager has

multiple choices across domains such as making capital investments, hiring

workers, introducing new products, and sales and marketing activities that he/she

is aware of and that are acceptable to corporate headquarters in China (e.g.

Acemoglu et al. 2007; Caza 2007; Chang and Wong 2003; Cheng et al. 2006;

Colombo and Delmastro 2004; Glaister et al. 2003; Marin and Verdier 2006).

In addition to the definition of discretion, managerial discretion theory contributes

to the understanding of the dimensionality of discretion, with early as well as recent

theoretical studies positing that managerial discretion consists of several types,

i.e. is multidimensional (e.g. Carpenter and Golden 1997, p. 195; Caza 2007, pp.

26–82; Chen 2006; Finkelstein and Peteraf 2007, p. 245; Hambrick and

Abrahamson 1995, p. 1439; Hambrick and Finkelstein 1987, pp. 371–402;

Hambrick et al. 1993, p. 409; see Sect. 2.1.2). This motivates the formulation of

Hypothesis 1 (dimensions of discretion), which was described in Sect. 2.1.2.

Second, the antecedents of managerial discretion (or causes, determinants or

sources of discretion) are introduced in managerial discretion theory as constructs

that determine the degree of managerial discretion obtained by a given manager.

For example, for the plant manager of the present study, a construct such as firm

size would act as an antecedent if plant managers in larger firms tended to

systematically possess different degrees of discretion than plant managers in

smaller firms. As noted above, the seminal work by Hambrick and Finkelstein

(1987, p. 369) qualitatively postulates that the managerial discretion of chief

executives varies ‘from very little to a great deal’ and that the particular extent of

discretion is determined by antecedents or ‘factors in the environment, the organi-

zation, and the chief executive’s own attributes’. More recent research on the

antecedents of managerial discretion has further extended these sources of discre-

tion on qualitative grounds. These resulting antecedents of discretion are divided

into three groups in the present study:

1. Environmental antecedents are factors that affect a manager’s discretion and

are common to all firms and managers in the environment. They correspond to

the antecedents of the ‘Task Environment’ in Hambrick and Finkelstein (1987,

pp. 378–389) for which they qualitatively suggest examples, such as product

differentiability, market growth, and industry concentration (e.g. Lieberson and

O’Connor 1972; Luce and Raiffa 1957; Porter 1980, p. 230). In the present

study, the environmental antecedent modelled is industry technology intensity

(e.g. Daniels 1993; Hatzichronoglou 1997; Loschky 2008; OECD 2005;

Onkelinx and Sleuwaegen 2010), which is defined at the industry level and

thus is common to all firms and managers in the industry.

2. Organisational antecedents are factors that affect a manager’s discretion and are

common to all managers in a particular firm. Hambrick and Finkelstein (1987,

pp. 378–389) label this category as ‘Internal Organization’, which is postulated

to include such antecedents as organisational size, age, and culture (e.g. Aldrich

1979; Cennamo et al. 2009; Galbraith 1967; Lodahl and Mitchell 1980; March


and Simon 1958; Mintzberg 1978; Peters and Waterman 1982; Reid 1968).49

The present study likewise models firm size as an organisational antecedent.

3. Managerial antecedents are factors that affect a manager’s discretion and are

particular to the manager under consideration. These antecedents are not included

in the present study, but could be modelled in future research. Originally,

Hambrick and Finkelstein (1987, pp. 378–389) proposed that individual ‘Mana-

gerial Characteristics’ may act as antecedents to discretion, such as a manager’s

aspiration level, commitment, cognitive processing ability, locus of control,

and power base as derived from e.g. tenure, shareholdings, and personal qualities

(e.g. Cyert andMarch 1963; March and Simon 1958; Miller et al. 1982; Peters and

Waterman 1982; Rotter 1966; Staw 1981). Subsequently, Finkelstein and Peteraf

(2007, pp. 237–243) added ‘Managerial Activities’ as a fourth antecedent of

discretion, which is classified under managerial antecedents for the reason that it

is particular to the manager under consideration. This fourth source of discretion is

an attempt to integrate principal-agent theory into managerial discretion theory,

and is further discussed below in Sect. 2.3.2 on principal-agent theory.

While managerial discretion theory has thus theoretically proposed a wide range

of potential antecedents of managerial discretion, scholars have rarely comprehen-

sively tested the realm of antecedents that have been put forward in the literature

over the past decades and instead have focused on quantitatively testing a subset of

antecedents. This is shown by the following examples in the aforementioned three

groups:

1. Environmental antecedents. For example, Hambrick and Abrahamson (1995)

and Finkelstein and Boyd (1998) confirm a subset of industry-level antecedents,

such as research and development (R&D) intensity, advertising intensity, market

growth, and concentration. Likewise, Cheng et al. (2006) find that managerial

discretion is increased by product differentiation, competitive product markets,

and higher skill intensity, and decreased by capital intensity and ownership at the

township level. Crossland (2007) adds national-level factors to the antecedents

of managerial discretion, namely cultural values, legal tradition, labour market

flexibility, and prevailing firm ownership structure. In line with the predictions

of these national-level or macro-environmental factors, Crossland and Hambrick

(2007) confirm that the effect of chief executive officers (CEOs) on company

performance is significantly higher (whether positive or negative) in the

United States than in Germany and Japan. Similarly, Bloom et al. (2009b) find

that factors such as trust, the rule of law, non-hierarchical religions, and product

market competition explain four-fifths of the cross-country variation in the

autonomy of plant managers within the firms investigated.

2. Organisational antecedents. While omitting the subsets of industry-level and

personality-related antecedents, Caza (2007, pp. 30–56) tests 13 other

49 In addition to such harder organisational antecedents as organisational size and age, softer

antecedents have been put forward in the literature. For example, Cennamo et al. (2009) contend

that top management may enlarge their managerial discretion by pursuing a broader stakeholder

management orientation.


antecedents that fall into the aforementioned groups and finds that the majority

of them are significant. What is most striking about Caza’s results is that the

observed relations of some of the antecedents to discretion differ by the dimen-

sion of discretion. For example, unit size was negatively related to ‘process

discretion’ and positively related to ‘resource discretion’. This again supports

the potential multidimensionality of the discretion construct, and was thus

considered in Sect. 2.1.2 in the formulation of Hypothesis 1. This potential

multidimensionality is further supported by Xiaoyang Li (2007, p. 16)—who

finds that a firm’s use of incentive compensation is positively associated with a

general manager’s labour autonomy but negatively with his/her investment

autonomy—and by Colombo and Delmastro (2004), who find that the degree

of discretion granted depends on the nature of the plant manager’s decision.

In addition, several other organisational antecedents have been empirically

confirmed, including firm size (Walters 1995; Zhang et al. 2006a), firm age

(Acemoglu et al. 2007; Xiaoyang Li 2007), and ownership structure (Colombo

and Delmastro 2004; Yougen Li and Zhao 2004).

3. Managerial antecedents. Individual managerial characteristics related to per-

sonality psychology have been successfully linked to perceived managerial

discretion, such as the locus of control, with Carpenter and Golden (1997,

p. 200) finding that ‘individuals with an internal locus of control tend to perceive

greater discretion than externals.’ Moreover, while omitting personality-related

antecedents, the results by Caza (2007, pp. 30–56) find that the effect of tenure as

predicting personal power was positive for ‘resource discretion’ and not signifi-

cant for ‘process discretion’. In contrast, formal knowledge and practical expe-

rience were positively linked to ‘process discretion’ but not significantly linked

to ‘resource discretion’ (again supporting the multidimensionality of discretion).

For similar managerial characteristics, Xiaoyang Li (2007, pp. 13–16) find that

general managers with longer tenure and more managerial experience tend to

have greater labour autonomy and greater investment autonomy. Walters (1995)

likewise empirically confirms top management tenure as an antecedent of

perceived managerial discretion. Whereas managerial antecedents relating to

‘Managerial Characteristics’ (Hambrick and Finkelstein 1987, pp. 378–389)

have thus been established, the testing of managerial antecedents relating to

‘Managerial Activities’ (Finkelstein and Peteraf 2007, pp. 237–243; see above)

largely remain open to further empirical investigation (which, however, lies

outside of the scope of the present study’s research objective).

Third, consequences of discretion have been postulated and empirically tested in

managerial discretion theory. In contrast to the research on the construct of discretion

(which addresses discretion itself) and the antecedents of discretion (which addresses

the constructs causing discretion), the research on the consequences of discretion

analyses relationships in which discretion can act as an independent variable, such

as the impact of managerial discretion on performance, which is investigated

in the present study. This stream of research has shown that managerial discretion

and similar constructs can exhibit significant relations with various potential

consequences, including managerial power (Carpenter and Golden 1997), managerial

compensation (Finkelstein and Boyd 1998; Magnan and St-Onge 1997; Rajagopalan


and Finkelstein 1992; Werner and Tosi 1995; Wright and Kroll 2002; Zhang and Xie

2008), workers’ incentives (Groves et al. 1994), the age of a successor CEO (Wang

2009), top management team tenure, trust (Perrone et al. 2003), strategic attention

(Abrahamson and Hambrick 1997), environmental commitment (Aragon-Correa

et al. 2004), pricing (Cameron 2000), organisational knowledge creation (Oh

2002), and research and development (Zhang et al. 2006a, b). Within this stream of

research, one may highlight a few studies that exemplify the potential importance of

managerial discretion for middle management and for China, given that this study’s

unit of analysis is a middle manager in China (see Chap. 3):

• Middle management. Perrone et al. (2003, p. 422) find that granting purchasing

managers more role autonomy (which they define as ‘the discretion that agents

have in interpreting and enacting their roles’) increases the trust that supplier

representatives place in purchasing managers. Oh (2002, pp. iii–iv) analyses

middle managers’ organisational knowledge creation and finds autonomy to be

the strongest predictor among six work environmental factors for the creation of

organisational knowledge.

• China. Moreover, the significance of managerial discretion in determining

outcomes has been documented in China. For instance, Zhang et al. (2006a, b)

find that in China, managerial discretion is associated with lower probabilities

and levels of R&D expenditure, and that it mediates the positive relationship

between firm scale and R&D expenditure.50

While the above review demonstrates that managerial discretion theory

contributes strongly to specifying the theoretical context for investigating the impact

of managerial discretion on performance (i.e. its construct, antecedents, and

consequences), it does not clearly specify the nature of the impact of discretion on

performance. The implications of managerial discretion theory for discretion’s

performance impact can be derived in relation to the discussion of the fictitious

scatter plot in Fig. 2.2 above (see Sect. 2.3.1.1). As already noted, Hambrick and

Finkelstein (1987, pp. 389–400) postulate that high managerial discretion (i.e. right-

hand side of Fig. 2.2) will generally lead to either very high or very low performance,

whereas low discretion (i.e. left-hand side of Fig. 2.2) will tend to be associated with

more stable and moderate performance. Consequently, managerial discretion theory

posits that the impact of a manager on performance is weak when discretion is low

(as predicted by population ecology) and strongwhen discretion is high (as predicted

by strategic choice theory; Hambrick and Finkelstein 1987). In other words, mana-

gerial discretion theory argues that when amanager has greater discretion, he/she can

potentially influence performance to a greater extent (whether positively or nega-

tively) than when he/she has lower latitude of action—leading to more performance

extremeness with greater discretion and more performance conformity with lower

50As defined by the research objective in Sect. 1.2, the dynamics of discretion (i.e. the analysis of

discretion over time, such as differentiating between short-term and long-term effects) are still at

an early stage of research in the literature (e.g. Finkelstein and Peteraf 2007, pp. 243–245;

Hutzschenreuter and Kleindienst 2007, p. 1; Kayhan 2008, pp. 1–6) and, in line with the present

study’s observational cross-sectional research design, are not addressed.


discretion (Finkelstein and Hambrick 1990; Misangyi 2002; Quigley and Hambrick

2009; Zhao et al. 2010).51 Indeed, using Lieberson and O’Connor’s (1972) variance

components analysis and further refinements (e.g.Mackey 2008), empirical research

has tested and supported this proposition that performance tends to be more variable

for firms with more discretion (e.g. Adams et al. 2005; Crossland and Hambrick

2007; Tang 2008). Nevertheless, while managerial discretion theory therefore

contends that as managerial discretion rises, the impact of managers on perfor-mance becomes stronger, it is left open whether this stronger impact is positive ornegative (or, as put by Crossland and Hambrick (2007, p. 2), the impact on perfor-

mance becomes stronger ‘for good and for ill’).

This postulated impact of discretion on performance in managerial discretion

theory is summarised by Fig. 2.3, which reinserts the fictitious scatter plot from

Fig. 2.2. When managerial discretion is low (i.e. on the left-hand side), the manager

has little latitude of action to influence performance, which is why performance is

expected to lie close to the average performance determined by environmental

circumstances (i.e. the horizontal axis), as in population ecology. As managerial

discretion rises, the manager has greater latitude of action to influence performance

(whether positively or negatively), which is why performance is expected to deviate

more strongly from average performance (i.e. the horizontal axis), as in strategic

choice theory. It follows (e.g. Adams et al. 2005; Crossland and Hambrick 2007;

Finkelstein and Hambrick 1990; Hambrick and Finkelstein 1987; Misangyi 2002;

Quigley and Hambrick 2009; Tang 2008; Zhao et al. 2010) that as managerial

discretion rises, performance could become more positive (i.e. the upwards-sloping

line) or more negative (i.e. the downwards-sloping line). As it is thus left open in

managerial discretion theory whether the relationship between discretion and

performance is positive or negative, the impact might be positive in some cases

and negative in other cases. If these cases were not separated, an empirical study

might not identify any significant impact of discretion on performance, leading to

the neutral impact indicated by the horizontal arrow in Fig. 2.3. As a result,

managerial discretion theory does not explicitly specify whether the impact of

discretion on performance is expected to be positive, negative or neutral. While

therefore being consistent with any empirically identified impact of discretion on

performance, managerial discretion theory does not make explicit predictions about

the nature of the impact of discretion on performance—and thus cannot satisfacto-

rily explain the contradictory empirical results of the discretion puzzle (see

Sect. 2.2). In terms of the empirical terminology introduced in Sect. 2.2, managerial

51 Finkelstein and Hambrick (1990, p. 488) define ‘the extent to which a firm’s performance is

similar to the average for the industry’ as performance conformity, with the opposite denoted by

performance extremeness (Quigley and Hambrick 2009, p. 3) or performance variability

(Misangyi 2002, pp. 36–37). For instance, for top management, chief executives with the discre-

tion to shape their company’s strategy might make radical changes to strategies and thereby impact

on performance more strongly than chief executives whose discretion is highly constrained, as in

regulated industries. For middle management, a plant manager whose discretion is constrained to

making capital investments of only say 1,000 RMB has less latitude to affect performance through

capital investments than a plant manager whose discretion is set to 1,000,000 RMB.


discretion theory therefore neither explicitly specifies whether the direct effect of

discretion on performance is positive, neutral or negative overall, nor focuses on

specifying moderating effects on this relationship, i.e. finding when the impact of

discretion on performance could be positive, neutral or negative.

In sum, managerial discretion theory was formally introduced by Hambrick and

Finkelstein (1987) as a bridge between diametrically opposed views of

organisational outcomes, namely population ecology and strategic choice theory.

While originally focusing on chief executives, the theory’s unit of analysis has been

extended from top management to middle management and is therefore applicable

to the present study, as noted in the literature review summary Table 2.8. On the one

hand, managerial discretion theory contributes strongly to fulfilling the present

study’s research objective by specifying the theoretical context for investigating

the impact of discretion on performance. In particular, it defines the construct of

managerial discretion, divides its causes into environmental, organisational, and

managerial antecedents, and relates discretion to possible consequences (including

performance)52 in a way that reconciles the polar views of population ecology and

High Performance

LowManagerialDiscretion

HighManagerialDiscretion

Low Performance

Fictitious performance-discretion data points for population ecology

Fictitious performance-discretion data points for strategic choice theory

Prediction ofPopulation Ecology

Prediction ofStrategic Choice Theory

Neutral

Fig. 2.3 Fictitious scatter plot for impact of discretion on performance

Source: The author’s own synthesis of managerial discretion theory, population ecology, and

strategic choice theory

52 In addition to performance, it has been demonstrated that discretion may exhibit significant

relations to such consequences as managerial power, managerial compensation, workers’

incentives, a successor chief executive officer’s age, top management team tenure, trust, strategic

attention, environmental commitment, pricing, organisational knowledge creation, and research

and development (see references above).


strategic choice theory. This theoretical context is the basis for developing the

study’s new discretion model in Sect. 4.1.1 and pivotal for formulating Hypothesis

1 (dimensions of discretion) in Sect. 2.1.2. On the other hand, however, while

managerial discretion theory specifies that discretion can have an impact on perfor-

mance in that as discretion rises, the impact of managers on performance becomes

stronger, it does not specify the nature of this impact on performance (i.e. whether

this stronger impact is positive or negative). Managerial discretion theory therefore

does not postulate mechanisms why the impact of discretion on performance could

Table 2.8 Literature review summary (Sect. 2.3.1 completed)

Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality







Impact on

performance:

• Direct

• Moderating

Contradictory

evidence

• Positive,

neutral,

negative

• Moderators

potentially

important

but rarely

modelled

(esp. firm

type and

firm size)

Theoretical

context specified

• Not specified

• Not specified.

Antecedents:

environmental,

organisational,

and managerial


management,

rarely China

Extended from

top to middle

management

Sections 2.3.2 and 2.3.3

Section 2.4

Hypothesis 1

Middle

management

in China

Hypothesis 2

Hypothesis 3


be positive or negative and in which way this impact may be moderated. This is

where the predictions of principal-agent theory and stewardship theory become

relevant, which are discussed next in Sects. 2.3.2 and 2.3.3, respectively.

2.3.2 Principal-Agent Theory

With a similar structure as in the previous section on managerial discretion theory,

the present section reviews the literature on principal-agent theory in two parts:

• Section 2.3.2.1 describes principal-agent theory briefly in terms of its historical

context, unit of analysis, assumptions, and predictions.

• Section 2.3.2.2 scrutinises principal-agent theory with respect to analysing

the impact of managerial discretion on performance. As in the previous section,

this entails a discussion of the construct of discretion (i.e. definition and

dimensionality), the antecedents of discretion, and the consequences of discre-

tion (i.e. direct effects and moderating effects).

2.3.2.1 Description of TheoryAs to its historical context, principal-agent theory (also termed agency theory) was

seminally developed by a number of scholars including Jensen and Meckling

(1976), Fama (1980), Spremann (1987), and Eisenhardt (1989) in order to analyse

situations in which a principal (e.g. an owner) has a contract that delegates

responsibility to an agent (e.g. a manager) to act on the principal’s behalf.

The theory’s historical roots date back further to the advent of the modern corpora-

tion, which created a divorce between the ownership and control of wealth

(Berle and Means 1932; Weidenbaum and Jensen 1993, p. 101; Zhang and Li

2008a, pp. 33–34). The capital requirements of the modern corporation typically

result in multiple owners, who become principals as they delegate responsibility to

agents (i.e. top management) for managing the firm, and this can result in

conflicting interests between owners and managers (Berle and Means 1932). For

this resulting contractual relation between owners (principals) and top management

(agents), the principal-agent theory has become the dominant paradigm in the

corporate governance literature (e.g. Davis et al. 1997b, p. 20; Dicke and Ott

2002, p. 464; Donaldson and Davis 1994, p. 151; Muth and Donaldson 1998,

p. 5; Shleifer and Vishny 1997, pp. 740–748; Thomsen and Pedersen 2000, p. 690).

In addition to being a dominant paradigm for top management as the unit ofanalysis, principal-agent theory has been widely applied to a variety of contractual

relations between shareholders, top management, middle management, and

employees in both the private and the public sector (e.g. Baysinger and Butler

1985; Berger et al. 1997; Bogart 1995; Brody 1996; Brush et al. 2000; Caza 2007;

2011; Chang and Wong 2003; Childs and Mauer 2008; Denis et al. 1997;

Dharwadkar et al. 2000; Lee and O’Neill 2003; McGubbins et al. 1987;

Ongore 2011; Van Slyke 2007; Walters 1995; Wood and Waterman 1991). This

broad application to differing hierarchical levels is consistent with the original

intention of principal-agent theory, as exemplified by Jensen and Meckling (1976,

p. 309), who write that ‘[t]he problem of inducing an “agent” to behave as if he were

maximizing the “principal’s” welfare is quite general. It exists in all organizations


and in all cooperative efforts—at every level of management’. As principal-agent

theory can therefore be utilised for studying contractual relations at various

organisational levels including the middle management level, it is applicable to

the present study’s unit of analysis, namely the plant manager in China (see

Chap. 3). In this study, top management at corporate headquarters in China delegate

responsibility for running a plant to the plant manager (see Fig. 1.1 in Sect. 1.3),

which makes top management the principal and the plant manager the agent in this

particular contractual relation.

Being a branch of new institutional economics,53 principal-agent theory is

typically characterised by two central assumptions (Eisenhardt 1989, pp. 58–59;

Levinthal 1988, p. 153; Spremann 1987, p. 3):54

• Assumption 1. The principal and the agent seek to individually maximise theirself-serving utility functions, which are mostly assumed to differ from each

other, thus introducing a conflict of interests (e.g. Albanese et al. 1997, p. 609;Chang andWong 2003, p. 25; Davis et al. 1997b, pp. 20–22; Dicke and Ott 2002,

p. 464; Eisenhardt 1989, pp. 58–59; Jensen and Meckling 1976; Jensen and

Murphy 1990, pp. 225–226; Levinthal 1988, p. 153; Spremann 1987, p. 3;

Thomsen and Pedersen 2000, p. 690; Van Slyke 2007, p. 162; Werner and

Tosi 1995, p. 1673). Applied to the present study, this assumption means that

both top management and the plant manager individually maximise their self-

serving utility functions, which, if assumed to differ, imply a goal conflict.

• Assumption 2. The principal cannot fully monitor the agent’s actions, meaning

there is post-contractual asymmetric information (i.e. hidden action or hidden

effort), since the agent has more information on his/her actions than the principal

(e.g. Eisenhardt 1989, p. 59; Jensen and Murphy 1990, p. 226; Khanchel 2009,

p. 97; Levinthal 1988, p. 153; Spremann 1987, p. 3; Van Slyke 2007, p. 162;

Werner and Tosi 1995, p. 1673). In the present study, this assumes that top

53New institutional economics, including principal-agent theory (see above) and transaction cost

economics (Argyres and Liebeskind 1999; Argyres and Mayer 2007; Bercovitz et al. 2006;

Coase 1937, 1960; Williamson 1985, 1991, 1996), evolved as a critique of neoclassical economics

by relaxing first-best assumptions, such as regarding information (i.e. asymmetric information in

principal-agent theory and information impactedness in transaction cost economics) or bounded

rationality (Pascha and Storz 2005, p. 16).54While the assumptions of principal-agent theory can differ in detail between differing models in

the literature (e.g. with assumptions at times also including bounded rationality and risk aversion),

assumptions 1 and 2 are central to principal-agent theory and tend to be common across differing

models (Eisenhardt 1989, pp. 58–59; Levinthal 1988, p. 153; Spremann 1987, p. 3). In addition to

these two central assumptions, Sect. 2.3.2.2 discusses two further assumptions commonly made by

principal-agent theorists when predicting the impact of managerial discretion on performance (i.e.

assumption 3: the agent’s managerial action is rational in terms of tending to produce the

outcomes that the agent intends; assumption 4: the agent’s interests are less aligned with

performance maximisation than the principal’s interests; e.g. Albanese et al. 1997, p. 610;

Chang and Wong 2003; Davis et al. 1997b, pp. 20–22; Eisenhardt 1989; Jensen 1986; Jensen

and Meckling 1976; Jensen and Murphy 1990; Shleifer and Vishny 1997; Spremann 1987;

Thomsen and Pedersen 2000; Werner and Tosi 1995, p. 1673).


management cannot fully monitor the plant manager’s managerial action.

In other words, top management cannot fully monitor how the plant manager

uses his/her ‘latitude of managerial action’ (i.e. managerial discretion;

Hambrick and Finkelstein 1987, pp. 371–378; see Sect. 2.1.1).

These assumptions allow principal-agent theory to make predictions about howthe agent uses his/her discretion relative to the interests of the principal.

The assumption that the principal cannot fully monitor the agent’s actions gives

the agent the discretion to pursue his/her self-serving interests to a certain extent

without the principal being able to tell (e.g. Fama and Jensen 1983b, p. 304;

Spremann 1987, p. 10; Zhao et al. 2010).55 If the utility functions that the principal

and agent are assumed to individually maximise diverge (i.e. if the agent’s interests

are not in line with the principal’s), then given this opportunity to pursue his/her self-

serving interests, the agent is predicted to necessarily act so as to maximise his/herown utility at the expense of the principal (Chang and Wong 2003, p. 6; Davis et al.

1997b, pp. 22–23). This predicted opportunism of the agent (i.e. self-interest seeking

with guile; Williamson 1975) is a consequence of the above assumptions of individ-

ually maximising self-serving utility functions given imperfect monitoring, as has

been demonstrated across a wide range of mathematical models in principal-agent

theory (e.g. Jensen and Meckling 1976; Levinthal 1988, p. 153; Spremann 1987;

Thepot 2007). Due to the agent’s opportunistic behaviour resulting from these

assumptions, agency costs arise to the principal in terms of losses to the principal’s

utility compared with the case where the agent acts in the principal’s best interest

(i.e. the agency problem; Baysinger and Butler 1985, p. 101; Berger et al. 1997,

pp. 1413–1414; Burkart et al. 1997, p. 705; Denis et al. 1997, p. 136; Eisenhardt

1989, p. 58; He et al. 2009, pp. 34–35; Jensen and Meckling 1976; Khanchel 2009,

p. 97; Lang et al. 1995, pp. 5–6; Xiaoyang Li 2007, p. 4; Shleifer and Vishny 1997,

pp. 740–741; Spremann 1987, p. 6; Xu et al. 2005, p. 4).

The intuition behind principal-agent theory’s predictions can be easily under-

stood with reference to the unit of analysis of the present study. The theory’s

assumptions then imply that top management at corporate headquarters in China

(principal) and the plant manager (agent) individually maximise their self-serving

utility functions, and that top management cannot fully monitor the plant manager’s

managerial action. Now suppose that the utility functions of top management and

the plant manager diverge, for example in relation to making capital investments.

Top management might be interested in making large capital investments only into

such machinery in the plant that is critical for the plant’s production processes to

run smoothly, whereas the plant manager might also favour large capital

investments into non-critical machinery on the grounds that he/she takes pleasure

55 For example, in Spremann’s (1987) principal-agent model total output and thus the principal’s

welfare are dependent on the agent’s hidden efforts as well as some exogenous risk. This implies

that the principal cannot fully distinguish between low effort on the part of the agent and bad luck

on the part of exogenous risk, so the asymmetric information (hidden effort) gives the agent

managerial discretion to pursue his/her own interests.


in running a modern factory. If instead of the assumption of imperfect monitoring

top management could fully monitor the plant manager’s capital investments, then

the plant manager would be constrained to act in top management’s best interest

and only buy critical machinery. If, however, the assumption of principal-agent

theory held that top management could not fully monitor the plant manager’s

actions, then rationally maximising his/her own utility, the plant manager would

conduct capital investments that he/she prefers (e.g. non-critical machinery) at the

expense of top management. Given the opportunity to follow his/her own interests

without being held accountable, principal-agent theory thus predicts the plant

manager to act in a self-serving way to the detriment of top management, which

produces the aforementioned agency costs.

In order for principals to protect themselves from the agent’s predicted self-

interest seeking (i.e. opportunism; Williamson 1975), principal-agent theory

prescribes that the principal impose internal control mechanisms upon the agent

that constrain the potential misconduct of the agents (Agrawal and Knoeber 1996;

Davis et al. 1997b, pp. 20–23; Eisenhardt 1989; Jensen and Meckling 1976;

Jensen and Ruback 1983; Khanchel 2009, pp. 97–99; Levinthal 1988, p. 153;

Van Slyke 2007, pp. 162–166).56 Various mechanisms have been designed for

curbing the agency costs to the principal, which relate to the aforementioned

assumptions of principal-agent theory, including compensation control mechanismsthat provide financial incentives and sanctions to align the interests of the agent and

the principal (see assumption 1; e.g. Agrawal and Knoeber 1996, p. 378;

Berger et al. 1997, p. 1411; Burkart et al. 1997, p. 705; Chang and Wong 2003,

p. 6; Cheng et al. 2006; Eisenhardt 1989, p. 60; Fama and Jensen 1983a, p. 345;

Jensen and Meckling 1976; Jensen and Murphy 1990, p. 226; Levinthal 1988,

p. 153; Spremann 1987, p. 10; Wang et al. 2008; Werner and Tosi 1995, p. 1673;

Zhang and Li 2008b), and monitoring control mechanisms such as audits and

performance evaluations to reduce the asymmetry of information (see assumption

2; e.g. Berger et al. 1997, p. 1411; Burkart et al. 1997, p. 705; Caza 2007; 2011;

Chang and Wong 2003, p. 6; Eisenhardt 1989, p. 60; Fama and Jensen 1983b;

Levinthal 1988, p. 153; Spremann 1987, pp. 10–11; Verhoest 2003, pp. 2–5;

Wang et al. 2008; Werner and Tosi 1995, p. 1673). For example, if top management

(principal) tied the plant manager’s compensation closer to his/her performance and

conducted performance evaluations, the plant manager might be more likely to

56Although the likelihood that the principal’s and agent’s interests diverge is substantial, the

interests may be aligned in some cases, whereby the agency problem is avoided and control

mechanisms become unnecessary (Davis et al. 1997b, p. 22). Nevertheless, as the interests of the

agent are difficult for the principal to judge ex ante (Williamson 1985), principal-agent theory

prudently recommends imposing control mechanisms to limit potential losses. Internal control

mechanisms are in general preferable to external control mechanisms, since external control

mechanisms such as acquisitions and divestures tend to come at a higher expense to the principal’s

utility (Walsh and Seward 1990, pp. 444–445). The extent to which agents fail to experience

discipline from this full range of control mechanisms has been termed entrenchment (Berger et al.1997, p. 1411).


undertake capital investments in the top management’s best interest. By prescribing

these and other control mechanisms, principal-agent theory has generated far-

reaching implications in a number of fields, most notably for the design of corporate

governance structures (see Shleifer and Vishny 1997).

2.3.2.2 Analysis of Impact of Discretion on PerformanceFollowing the review of principal-agent theory in general terms (e.g. historical

context, unit of analysis, assumptions, and predictions), the theory is now explored

in the light of the present study’s research objective of investigating the impact of

managerial discretion on performance. It is shown that in contrast to managerial

discretion theory (see Sect. 2.3.1), principal-agent theory does not fully specify the

theoretical context of managerial discretion in terms of defining the construct of

discretion, its antecedents, and its consequences. Instead, principal-agent theory

postulates mechanisms that help explain why the impact of discretion on perfor-

mance could be of a particular nature and in what ways it might be moderated—all

of which is important for formulating the study’s hypotheses. As in the previous

section on managerial discretion theory, this discussion is carried out in the

following sequence:



3. Consequences of managerial discretion (i.e. direct and moderating effects).

First, in contrast to managerial discretion theory (see Sect. 2.3.1), principal-

agent theory has paid little attention to defining the construct of managerial

discretion, as described in Sect. 2.1. Principal-agent theorists have tended not to

explicitly define managerial discretion, thereby continuing a tendency of new

institutional economics (e.g. Burkart et al. 1997; Childs and Mauer 2008; He

et al. 2009; Lang et al. 1995; Williamson 1963). Nevertheless, the meaning of

managerial discretion in these studies does tend to concur with that of managerial

discretion theory (e.g. Chang andWong 2003, p. 2; Hambrick and Finkelstein 1987,

pp. 371–378; Khanchel 2009, p. 97; Shleifer and Vishny 1997, p. 742; Spremann

1987, p. 10; see Sect. 2.1.1). As to the dimensionality of discretion, principal-agent

theory has often modelled discretion as a single (unidimensional) construct with a

single particular impact on performance (e.g. Chang and Wong 2003; He et al.

2009; Spremann 1987, p. 18; Xu et al. 2005), which contrasts with the multidimen-

sionality postulated in managerial discretion theory (e.g. Carpenter and Golden

1997, p. 195; Caza 2007, pp. 26–82; Huiyuan Chen 2006; Finkelstein and Peteraf

2007, p. 245; Hambrick and Abrahamson 1995, p. 1439; Hambrick and Finkelstein

1987, pp. 371–402; Hambrick et al. 1993, p. 409; see Sect. 2.1.2). It is this

discontinuity between the unidimensionality and multidimensionality of manage-

rial discretion that motivates the present study to test Hypothesis 1 (dimensions of

discretion), as explained in Sect. 2.1.2.

Second, although principal-agent theory does not explicitly discuss the

antecedents of discretion (or causes, determinants or sources of discretion; see

Sect. 2.3.1), its aforementioned theoretical content does have implications for the

antecedents. These implications can be derived by revisiting the definition of mana-

gerial discretion (Abrahamson and Hambrick 1997, p. 513; Carpenter and Golden


1997, p. 187; Caza 2007, p. 27; Chang and Wong 2003, p. 2; Crossland 2007, p. 1;

Crossland and Hambrick 2007, p. 767; Finkelstein and Boyd 1998, p. 179;

Finkelstein and Hambrick 1990, p. 484; Hambrick and Abrahamson 1995, p. 1427;

Rajagopalan and Finkelstein 1992, p. 32; see Sect. 2.1.1), whereby discretion is

defined as the extent to which a manager (i.e. agent) has multiple courses of action

across various domains of his/her work that he/she is aware of and that are acceptable

to the parties that possess power to constrain the manager (i.e. principal). Multiple

courses of action thus count towards the plant manager’s (agent’s) discretion only if

they ‘lie within the zone of acceptance of powerful parties’ (principal), which in this

study is top management in China (Hambrick and Finkelstein 1987, p. 378).

Principal-agent theory’s assumption of asymmetric information (i.e. the inability of

the principal to fully monitor the agent’s actions; see assumption 2 above; e.g.

Eisenhardt 1989, p. 59; Jensen and Murphy 1990, p. 226; Khanchel 2009, p. 97;

Levinthal 1988, p. 153; Spremann 1987, p. 3; Van Slyke 2007, p. 162; Werner and

Tosi 1995, p. 1673) widens the zone of acceptance of the principal, since to the extent

that the principal cannot monitor the agent’s activities and thus hold him/her

accountable, the principal is bound to accept the agent’s actions (e.g. Spremann

1987, p. 10).57 The more difficult it is for the principal to monitor the agent’s

activities, the more latitude (i.e. discretion) the agent has in choosing additional

courses of action, even if they are against the principal’s best interests. Consequently,

the asymmetry of information (i.e. the inability of the principal to fully monitor the

agent’s actions) gives the agent additional managerial discretion, which makes it a

potential antecedent of managerial discretion.

A similar argument is made by Finkelstein and Peteraf (2007, pp. 237–243) in

their combination of managerial discretion theory and principal-agent theory

(which was mentioned in Sects. 1.1, 2.1.1, and 2.3.1 above). They postulate that

different characteristics of managerial activities affect the ability of key

stakeholders (i.e. powerful parties) to pre-specify and monitor the manager’s

work and thereby create or constrain the manager’s discretion. In particular, the

complexity, uncertainty, and lack of observability (e.g. Eisenhardt 1989;

Holmstrom 1979; Jensen and Murphy 1990; Rumelt 1984; Tirole 1988) of mana-

gerial activities are all expected to intensify the asymmetry of information between

the principal and the agent, which, as explained based on principal-agent theory

above, can increase the manager’s discretion. Finkelstein and Peteraf (2007,

pp. 237–243) therefore view the complexity, uncertainty, and lack of observability

of managerial activities as sources of discretion, which is why ‘Managerial

Activities’ were integrated as the fourth type of antecedents in Sect. 2.3.1.2.

The present study’s new theoretical discretion model developed in Sect. 4.1.2

integrates Finkelstein and Peteraf’s (2007, pp. 237–243) ‘Managerial Activities’-based

57As noted above, in Spremann’s (1987) principal-agent model, low output (i.e. low performance)

can be caused by either low effort on the part of the agent or bad luck due to exogenous risk. As the

principal cannot fully monitor the agent’s effort and thus cannot fully hold the agent accountable

for his/her actions, this asymmetric information (hidden effort) gives the agent managerial

discretion to pursue his/her own interests.


synthesis of managerial discretion theory and principal-agent theory and furthermore

attempts to advance their combination of these two theories in a number of ways:

• Not only is the nature of managerial activities (i.e. complexity, uncertainty, and

lack of observability) taken into account as an antecedent of discretion, but also

the aforementioned monitoring control mechanisms of principal-agent theory.

As explained above, Finkelstein and Peteraf (2007, pp. 237–243) argue that as

the agent’s managerial activities become more complex, more uncertain, and

less observable, the inability of the principal to monitor the agent’s activities (i.e.

asymmetry of information) becomes more severe and therefore causes greater

managerial discretion. As this postulated causality runs via the above assump-

tion of incomplete monitoring (i.e. asymmetric information), anything else that

affects the principal’s ability to monitor the agent might likewise act as an

antecedent of discretion. In particular, monitoring control mechanisms that

potentially mitigate the asymmetry of information, such as audits and perfor-

mance evaluations (e.g. Berger et al. 1997, p. 1411; Burkart et al. 1997, p. 705;

Caza 2007; 2011; Chang andWong 2003, p. 6; Eisenhardt 1989, p. 60; Fama and

Jensen 1983b; Levinthal 1988, p. 153; Spremann 1987, pp. 10–11; Verhoest

2003, pp. 2–5; Wang et al. 2008; Werner and Tosi 1995, p. 1673), could also

constrain the agent’s managerial discretion. Therefore, monitoring control

mechanisms are embedded in the present study’s theoretical discretion model

(see Sect. 4.1.2) as antecedents, along with the nature of managerial activities.58

• In addition to integrating principal-agent theory in terms of explaining when a

manager may have lower or greater discretion (i.e. as antecedents of discretion),this study also integrates principal-agent theory in terms of explaining whether

lower or greater discretion improves or reduces performance (i.e. in terms of the

consequences of discretion). This extension is discussed below in terms of both

direct effects and moderating effects (e.g. Baron and Kenny 1986, p. 1174;

Henseler and Fassott 2010, p. 714).

• Finally, the present study’s approach integrates stewardship theory (e.g.

Albanese et al. 1997; Arthurs and Busenitz 2003; Corbetta and Salvato 2004;

Davis et al. 1997a, b; Dicke and Ott 2002; Donaldson 1990; Donaldson and

58 For example, suppose top management at corporate headquarters in China (principal) could

initially not fully monitor the plant manager’s (agent’s) managerial activities, such as his/her

capital investments, perhaps due to their complex, uncertain or unobservable nature. The plant

manager might then have the discretion to undertake certain capital investments out of self-serving

interests and against the interests of top management. The nature of the plant manager’s manage-rial activitieswould thus constitute an antecedent of discretion. Now suppose that top management

imposed stricter monitoring control mechanisms, such as audits and performance evaluations that

made it easier to monitor the plant manager’s activities. Top management would then be more able

to hold the plant manager accountable for the effects of his/her capital investments and not every

capital investment would fall into their zone of acceptance (Hambrick and Finkelstein 1987,

p. 378). The monitoring control mechanisms would thus reduce the choices available to the

plant manager for making investments that are acceptable to top management in China—i.e.

they would reduce the manager’s discretion.


Davis 1989, 1991, 1993, 1994, p. 159; Eddleston and Kellermanns 2007;

Fox and Hamilton 1994; Lane et al. 1999; Liu and Cai 2004; Miller and

Le Breton-Miller 2006; Mills and Keast 2009; Muth and Donaldson 1998;

Salvato 2002; Tian and Lau 2001; Tosi et al. 2003; Van Slyke 2007;

Vargas Sanchez 2001, 2004, 2005; Zahra 2003; see Sect. 2.3.3) as well as the

potential multidimensionality of managerial discretion (e.g. Barnabas and

Mekoth 2010; Carpenter and Golden 1997, p. 195; Caza 2007, pp. 26–82;

Chen 2006; Finkelstein and Peteraf 2007, p. 245; Groves et al. 1994, p. 190;

Hambrick and Abrahamson 1995, p. 1439; Hambrick and Finkelstein 1987,

pp. 371–402; Hambrick et al. 1993, p. 409; March and Simon 1958;

Perrone et al. 2003, pp. 422–423; see Sect. 2.1.2) into its empirically-verifiable

model. This again sets the present study apart from existing approaches in the

literature that combine managerial discretion theory and principal-agent

theory—both those approaches that have begun to integrate principal-agent

theory into the antecedents of managerial discretion (Finkelstein and Peteraf

2007, pp. 237–243) and those that have begun to integrate principal-agent theory

into the consequences of managerial discretion (e.g. Caza 2007; 2011).

Third, as to the consequences of discretion (i.e. direct and moderating effects),

the preceding discussion in Sect. 2.3.1.2 has shown that managerial discretion

theory does not clearly specify the nature of the impact of discretion on perfor-

mance. Managerial discretion theory merely contends that discretion can have an

impact on performance in that as discretion rises, the impact of managers on

performance may become stronger (e.g. Adams et al. 2005; Crossland and

Hambrick 2007; Finkelstein and Hambrick 1990; Hambrick and Finkelstein 1987;

Misangyi 2002; Quigley and Hambrick 2009; Tang 2008; Zhao et al. 2010).

Specifically, as discretion rises, the manager has greater latitude of action to

influence performance (whether positively or negatively), which is why perfor-

mance is expected to deviate more strongly from average performance (i.e. the

horizontal axis in Fig. 2.3 in Sect. 2.3.1.2) when discretion is higher. Yet manage-

rial discretion theory does not postulate mechanisms that explain why the stronger

impact of discretion on performance could be positive or negative and in which way

it might be moderated.

By contrast, principal-agent theory makes predictions for both the overall

expected direction of the impact of managerial discretion on performance (i.e.

direct effect) and the potential moderators on this relationship (i.e. moderating

effects). These result in part from the theory’s two aforementioned central

assumptions—namely assumption 1 (individually maximising self-serving utility

functions) and assumption 2 (imperfect monitoring; e.g. Albanese et al. 1997,

p. 609; Chang and Wong 2003, p. 25; Davis et al. 1997b, pp. 20–22; Dicke and

Ott 2002, p. 464; Eisenhardt 1989, pp. 58–59; Jensen and Meckling 1976;

Jensen and Murphy 1990, pp. 225–226; Khanchel 2009, p. 97; Levinthal 1988,

p. 153; Spremann 1987, p. 3; Thomsen and Pedersen 2000, p. 690; Van Slyke 2007,

p. 162; Werner and Tosi 1995, p. 1673)—as well as from two further assumptions

commonly made by principal-agent theorists (see below). As explained above,

assumption 2, that of imperfect monitoring, gives the agent the discretion to pursue

self-serving interests to a certain extent without the principal being able to tell


(e.g. Fama and Jensen 1983b, p. 304; Spremann 1987, p. 10; Zhao et al. 2010).

Moreover, the assumption 1 of individually maximising self-serving utility

functions implies that given this opportunity to pursue self-serving interests, the

agent necessarily acts so as to maximise his/her own utility, opportunistically

pursuing self-serving interests even if this comes at a cost to the principal (e.g.

Chang and Wong 2003, p. 6; Davis et al. 1997b, pp. 22–23; Jensen and Meckling

1976; Levinthal 1988, p. 153; Spremann 1987; Thepot 2007). It follows that underimperfect monitoring, discretion gives the agent the opportunity to pursue self-serving interests at the expense of the principal, which the agent chooses to do sincethis maximises his/her self-serving utility function according to principal-agenttheory. Under these two assumptions, increasing discretion thus gives the agent

greater leeway (i.e. more multiple courses of action; see Hambrick and Finkelstein

1987, pp. 371–378 in Sect. 2.1.1) to pursue self-serving interests instead of striving

to work towards the principal’s interests, and the agent chooses to use this additional

discretion in a way that he/she expects to better fulfil his/her own interests rather than

those of the principal, which results in agency costs (e.g. Baysinger and Butler 1985,

p. 101; Berger et al. 1997, pp. 1413–1414; Burkart et al. 1997, p. 705; Denis et al.

1997, p. 136; Eisenhardt 1989, p. 58; He et al. 2009, pp. 34–35; Jensen and Meckling

1976; Khanchel 2009, p. 97; Lang et al. 1995, pp. 5–6; Xiaoyang Li 2007, p. 4;

Shleifer and Vishny 1997, pp. 740–741; Spremann 1987, p. 6; Xu et al. 2005, p. 4).

Consequently, principal-agent theory predicts that the agent uses additional mana-gerial discretion to produce outcomes that he/she intends to be closer to fulfilling his/her own interests than without the additional discretion.

In order to move from this prediction of principal-agent theory (derived above

from assumptions 1 and 2) to predictions regarding the impact of managerial

discretion on performance, two further assumptions are required:

• Assumption 3. It is necessary to assume that the agent’s managerial action isrational in terms of tending to produce the outcomes that the agent intends. With

this assumption, principal-agent theory predicts that the agent uses additional

managerial discretion to produce outcomes that are (rather than that he/she

intends to be) generally closer to fulfilling his/her own interests than without

the additional discretion. In other words, the agent is assumed to be rational,

having the ability, knowledge, and information that are necessary to allow his/

her managerial action to translate into the intended outcomes. One may view this

assumption as implicit in assumption 1, as individual self-serving utility

maximisation posits that the agent is a rational homo-economicus (e.g. Albanese

et al. 1997, p. 610; Davis et al. 1997b, pp. 20–22; Werner and Tosi 1995,

p. 1673). Nevertheless, self-interest and rational behaviour have also been

treated as separate assumptions in new institutional economics (e.g. Williamson

1963, p. 1054). Moreover, making assumption 3 explicit allows this section to

define more precisely the theoretical mechanisms by which discretion can affect

performance, which in turn allows this section to define further moderating

effects on the impact of discretion on performance (see below).

• Assumption 4. In order for principal-agent theory to predict whether granting

discretion increases, does not alter or decreases performance, the nature of the

agent’s and the principal’s individual self-serving utility functions from


assumption 1 need to be further specified as to whether the agent’s interests or theprincipal’s interests are more aligned with maximising performance. With

assumptions 1 to 3 only, principal-agent theory merely predicts that additional

discretion produces outcomes that are closer to fulfilling the agent’s interests and

further away from fulfilling the principal’s interests (see above), but not whether

these outcomes entail higher or lower performance. As is explained in detail in

Box 2.1, depending on whose interests are more performance-maximising, addi-

tional managerial discretion might increase, not alter or decrease performance.

Nevertheless, Chang and Wong (2003, p. 7) explain that principal-agent theory

generally assumes that the agent is less profit-maximising than the principal, i.e.

‘[a]gency theory assumes that controlling parties’ objective is to maximize profits

but that managers have non-profit-maximizing objectives.’ Indeed, the assumption

that agents have non-profit-maximising objectives is prevalent in the principal-

agent literature (e.g. Chang and Wong 2003; Eisenhardt 1989; Jensen 1986;

Jensen and Meckling 1976; Jensen and Murphy 1990; Shleifer and Vishny

1997; Spremann 1987) and the possibility that principals (i.e. controlling parties)

‘have objectives other than profit maximization is rarely considered in this litera-

ture’ (Thomsen and Pedersen 2000, p. 690). It follows that in most cases,

principal-agent theorists contend that the agent’s interests are less performance-maximising than the principal’s interests. Under this assumption (i.e. assumption

4), the aforementioned prediction that additional discretion yields outcomes closer

to fulfilling the agent’s interests and further away from the principal’s interests

implies that these outcomes resulting from additional discretion engender lower

rather than higher performance (see Box 2.1).

Box 2.1: Agent’s and Principal’s Interests in Maximising Performance

This Box scrutinises the aforementioned assumption 4 of principal-agent

theory in greater detail, demonstrating that assumption 4 is critical for the

theory to predict a negative impact of managerial discretion on performance.

Specifically, whether the agent’s interests or the principal’s interests are more

aligned with maximising performance drastically alters the predicted impact

of discretion on performance in the following way:

1. If the agent’s interests are less performance-maximising than theprincipal’s, then additional managerial discretion is predicted to decreaseperformance (d < 0). The reason is that when the agent uses the additional

discretion to pursue his/her own interests at the expense of the principal,

outcomes approach the agent’s interests and thereby drift further away

from performance-maximisation.59 Hence, when the principal is more

performance-inclined than the agent, the predicted direct effect of discretion

59 For example, when top management at corporate headquarters in China (i.e. the principal) is

more interested in high performance than the plant manager (i.e. the agent), then the plant manager

might use the additional discretion to make capital investments that suit his/her personal agenda at

the expense of performance.


on performance is negative (d < 0 in Equations (2.1) and (2.2) in Sect. 2.2.1

above). This is the case most often assumed in principal-agent theory (as

recorded in assumption 4 above; e.g. Chang and Wong 2003, pp. 2–7;

Eisenhardt 1989; Jensen 1986; Jensen and Meckling 1976; Jensen and

Murphy 1990; Shleifer and Vishny 1997; Spremann 1987; Thomsen and

Pedersen 2000, p. 690) and thus principal-agent theory tends to predict anegative impact of managerial discretion on performance (e.g. Caza 2007,p. 10; Caza 2011; Chang and Wong 2003, p. 7; Davis et al. 1997b, p. 38;

Hutzschenreuter and Kleindienst 2007, p. 4; Jensen and Murphy 1990;

Zhao et al. 2010). As noted in Sect. 2.2, various scholars have found

significant evidence that the direct effect of discretion on performance is

negative (d < 0), which has been interpreted as evidence in support of

principal-agent theory (e.g. He et al. 2009; Heinecke 2011; Stano 1976;

Williamson 1963; Xu et al. 2005).

2. If the agent’s interests are more performance-maximising than theprincipal’s, then additional managerial discretion is predicted to increaseperformance (d > 0). Although the agent still abuses discretion in pursuit

of his/her own interests and this moves outcomes further away from those

desired by the principal, outcomes exhibit higher performance in line with

the agent’s interests. In the present study, this would assume that the plant

manager is more interested that he/she achieves high performance in his/

her plant than top management at corporate headquarters in China is

concerned with performance in the plant. While this assumption may

seem unlikely for this unit of analysis, some scholars have interpreted

an empirically positive effect of discretion on performance (d > 0) as

confirming this assumption. For instance, Chang and Wong (2003, 2004)

find a positive relationship between top management discretion and per-

formance in China and attribute this on qualitative grounds to managers

having relatively more profit-maximising objectives than the mostly state-

owned controlling parties in China. Nevertheless, Chang and Wong (2003,

pp. 1–7) admit that while this positive impact is a possibility, principal-

agent theory generally predicts a negative impact of discretion on

performance by assuming that the principal is mostly more performance-

maximising than the agent (assumption 4; e.g. Chang and Wong 2003,

p. 2; Thomsen and Pedersen 2000, p. 690).

3. If the agent’s interests are equally performance-maximising as theprincipal’s, then additional managerial discretion is predicted not toalter performance through the aforementioned theoretical mechanism inprincipal-agent theory (d � 0).60 This case is, however, unlikely due to

(continued)

60 Even when accepting a close alignment of the principal’s and agent’s interests, the impact of

discretion on performance may not be insignificant due to the additional theoretical mechanism

described in Sect. 2.3.3.


assumption 1 of principal-agent theory that the principal and the agent seek

to individually maximise their self-serving utility functions, since these

utility functions are mostly expected to diverge rather than be aligned (e.g.

Albanese et al. 1997, p. 609; Chang and Wong 2003, p. 25; Davis et al.

1997b, pp. 20–22; Dicke and Ott 2002, p. 464; Eisenhardt 1989, pp. 58–59;

Jensen and Meckling 1976; Jensen and Murphy 1990, pp. 225–226;

Levinthal 1988, p. 153; Spremann 1987, p. 3; Thomsen and Pedersen

2000, p. 690; Van Slyke 2007, p. 162; Werner and Tosi 1995, p. 1673).

An exception occurs when control mechanisms (see Sect. 2.3.2.1) are so

effective that they unhinge the assumptions of principal agent theory, i.e.

when compensation control mechanisms sufficiently align the agent’s

interest with performance maximisation and monitoring control

mechanisms sufficiently mitigate the principal’s asymmetry of informa-

tion. In this exceptional case, principal-agent theory can be consistent with

an insignificant impact of discretion on performance (d � 0) (e.g. Agrawal

and Knoeber 1996, p. 377; Chang and Wong 2003, p. 7). In most cases,

however, principal-agent theory makes assumptions 1 to 4 and therefore

predicts a negative impact of discretion on performance (d < 0) rather than

the insignificant impact (d � 0) found at times in the empirical literature

(see Sect. 2.2.2; e.g. Caza 2011; Groves et al. 1994; Yougen Li and Zhao

2004; Lopez-Navarro and Camison-Zornoza 2003; Venaik 1999).

It follows that in the principal-agent framework, assumption 4 is critical

for determining whether discretion decreases (d < 0), increases (d > 0) or

does not alter (d � 0) performance. This suggests that a potential explanation

of the discretion puzzle (see Sects. 1.1 and 2.2.2) is that studies finding a

positive impact of discretion on performance have focused on agents with

more performance-maximising interests than their principals, studies finding

a negative impact have focused on agents with less performance-maximising

interests, and studies finding an insignificant impact have focused on agents

with interests similar to those of their principals (or an offsetting number of

both types of agents). Although this explanation is limited in that it requires

the assumption that agents are more performance-maximising than their

principals in order to predict a positive effect (d > 0), it is used as a starting

point for developing a more compelling explanation in Sect. 2.3.3.61

It follows that when assumptions 1 to 4 are made in line with principal-agent

theory, the theory predicts a negative impact of managerial discretion on perfor-mance, since it postulates that managers (agents) use discretion to pursue their own

61As explained in Sect. 2.3.3, stewardship theory proposes an additional theoretical mechanism

which can help explain a positive impact of discretion on performance. After discussing this

mechanism, Sect. 2.3.3 correspondingly extends this potential explanation of the contradictory

evidence of the discretion puzzle.


interests at the cost of the principal, which under assumption 4 tends to work to the

detriment of performance (see Box 2.1). The literature on principal-agent theory

therefore generally contends that managerial discretion has a negative direct effecton performance (e.g. Caza 2007, p. 10; Caza 2011; Chang and Wong 2003, p. 7;

Davis et al. 1997b, p. 38; Hutzschenreuter and Kleindienst 2007, p. 4; Jensen and

Murphy 1990; Zhao et al. 2010).62 This prediction of principal-agent theory that

discretion tends to decrease performance (d < 0) contributes to the formulation of

the present study’s final hypothesis, Hypothesis 4, which is discussed in Sect. 2.4.3.

Before turning from the predicted negative direct effect of discretion on perfor-

mance to moderating effects, the above discussion is consolidated in Fig. 2.4 with

reference to the present study’s unit of analysis. In accordance with the preceding

fictitious scatter plots (see Fig. 2.1 in Sect. 2.2 and Figs. 2.2 and 2.3 in Sect. 2.3.1),

Fig. 2.4 measures managerial discretion (D) on the horizontal axis and performance

(P) on the vertical axis. The downwards-sloping line depicts performance (P) as afunction of discretion (D ) under principal-agent theory, with the negative slope

corresponding to the aforementioned prediction that discretion negatively affects



HighManagerial

Discretion (D)

Low Performance (P)

Moderating Effects (m)Agent's Natural Predisposition to Performance MaximisationCompensation Control MechanismsNature of Managerial ActivitiesMonitoring Control MechanismsAgent's Ability, Knowledge, and Information


Source: Principal-agent theory

62 As explained in Box 2.1, it is possible to assume that the principal is less performance-



theory (Thomsen and Pedersen 2000, p. 690). Nevertheless, it should be noted that the principal-

agent literature recognises that maintaining a certain level of performance might be in the

manager’s own best interest due to ‘both the discipline and opportunities provided by the markets

for their services, both within and outside of the firm’ (Fama 1980, p. 289).


performance (d < 0).63 The reasoning why principal-agent theory predicts discre-

tion to decrease performance (i.e. a downwards-sloping line) is summarised as

follows:

• Managerial discretion theory (see Sect. 2.3.1) implies that as discretion rises

(moving rightwards on Fig. 2.4), the manager has greater latitude of action to

influence performance (whether positively or negatively), which is why perfor-

mance is then expected to deviate more strongly from average performance (i.e. lie

further above or below the horizontal axis in Fig. 2.4; e.g. Adams et al. 2005;

Crossland and Hambrick 2007; Finkelstein and Hambrick 1990; Hambrick and

Finkelstein 1987; Misangyi 2002; Quigley and Hambrick 2009; Tang 2008;

Zhao et al. 2010). With this prediction of managerial discretion theory, it is,

however, left open whether the line depicted in Fig. 2.4 is upwards-sloping (i.e.

positive;d > 0) or downwards-sloping (i.e. negative;d < 0). For example, raising a

plant manager’s discretion for making capital investments from zero RMB (left in

Fig. 2.4) to 100,000RMB (right in Fig. 2.4) gives the plantmanager the opportunity

to make investments in a way that might either positively (‘good investments’) or

negatively (‘bad investments’) affect performance.

• Principal-agent theory then introduces assumptions 1 to 4 discussed above,

which collectively imply that granting a manager additional discretion tends to

reduce performance (or that in the above example, a plant manager with capital

investment discretion tends tomake ‘bad investments’). As the agent individually

maximises his/her self-serving utility function (assumption 1), under imperfect

monitoring (assumption 2), discretion gives the agent the opportunity to pursue

self-serving interests at the expense of the principal, which the agent chooses to do

since this maximises his/her utility function (e.g. Albanese et al. 1997, p. 609;

Chang andWong 2003, p. 25; Davis et al. 1997b, pp. 20–22; Dicke and Ott 2002,

p. 464; Eisenhardt 1989, pp. 58–59; Jensen and Meckling 1976; Jensen and

Murphy 1990, pp. 225–226; Khanchel 2009, p. 97; Levinthal 1988, p. 153;

Spremann 1987, p. 3; Thomsen and Pedersen 2000, p. 690; Van Slyke 2007,

p. 162; Werner and Tosi 1995, p. 1673). Provided that the agent’s managerial

action tends to produce the outcomes that the agent intends (assumption 3),

greater discretion is thus predicted to be abused by the agent in a way that shifts

outcomes closer to fulfilling the agent’s self-serving interests albeit at the cost of

the principal’s interests (e.g. Albanese et al. 1997, p. 610; Chang andWong 2003,

p. 6; Davis et al. 1997b, pp. 20–23; Jensen and Meckling 1976; Levinthal 1988,

p. 153; Spremann 1987; Thepot 2007; Werner and Tosi 1995, p. 1673). Finally,

when assuming that the agent’s interests are less aligned with maximising perfor-

mance than the principal’s interests (assumption 4), then the greater discretion

that the agent uses opportunistically to get closer to his/her self-serving interests is

predicted to entail a reduction in performance (e.g. Caza 2007, p. 10; Caza 2011;

63 As noted above, Sect. 2.4.3 explains that this prediction of principal-agent theory (d < 0) is

recorded as the principal-agent hypothesis for Hypothesis 4, which is indicated on the downwards-

sloping line in Fig. 2.4.


Chang and Wong 2003, p. 7; Davis et al. 1997b, p. 38; Hutzschenreuter and

Kleindienst 2007, p. 4; Jensen and Murphy 1990; Zhao et al. 2010). In terms of

Fig. 2.4, principal-agent theory then predicts that granting the plant manager

additional discretion such as for making capital investments will mostly lead to a

reduction in performance (d < 0), because the plantmanagermaximises utility by

pursuing his/her self-serving interests (assumption 1) and given that top manage-

ment at corporate headquarters in China cannot fully monitor the plant manager’s

actions (assumption 2), he/she will choose to make such capital investments that

produce the plant manager’s desired outcomes (assumption 3) but come at the

expense of performance (assumption 4).64 Hence, principal-agent theory predicts

the plant manager to use additional discretion (i.e. moving rightwards in Fig. 2.4)

to pursue his/her self-serving interests (e.g. ‘bad investments’ for personal

reasons), which tends to reduce performance (i.e. as represented by the

downwards-sloping line in Fig. 2.4).

Finally, in addition to the predicted negative overall direct effect of discretion onperformance (d < 0), a number of moderating effects can be derived from the

preceding review of principal-agent theory. Moderating effects recognise that the

impact of discretion on performance may lie above or below the generally negative

direct effect (d) in certain situations, depending on the values of moderator variables

(e.g. Baron and Kenny 1986, p. 1174; Henseler and Fassott 2010, p. 714). In terms of

Fig. 2.4, although the predicted impact of discretion on performance is generally

negative,moderating effects allow this impact to bemore or less negative in different

situations, pivoting the line to become more or less downwards-sloping. Intuitively,

the moderator variables denote influences on managers that affect the way in which

managers use their discretion and therefore how the managers’ actions impact on

performance.65 By deriving the influences (i.e. moderators) that determine whether

64 As noted above, if one instead assumes that the agent’s (i.e. plant manager’s) interests are more

aligned with maximising performance than those of the principal (i.e. top management at corporate

headquarters in China), then granting discretion might increase performance (e.g. Chang and

Wong 2003). The plant manager would still use discretion opportunistically to produce the

outcomes that he/she desired, but when assuming that these agent-desired outcomes are more

performance-maximising than those of the principal, then this self-interest seeking behaviour of

the agent would increase performance (albeit to the discontent of the principal).65Moderating effectswere formally introduced in Sect. 2.2.1 and are discussed at length in Sect. 4.3. It

was shown that the relationship between discretion and performance is oftenmodelled by an equation

similar toP ¼ d � Dþ c � Cþ m � D �Mð Þ þ ε(seeEquation (2.1) inSect. 2.2.1),which is depicted bythe downwards-sloping line in Fig. 2.4 (when the control variableC and the moderator variableM are

normalised to zero). By partial differentiation, the total impact of discretion on performance was

derived as @P @D= ¼ d þ m �M (see Equation (2.2) in Sect. 2.2.1), i.e. the sum of the direct effect ofdiscretion on performance (d) and the moderating effect (m) multiplied by the moderator variable’s

value (M) (see Sect. 4.3.2 on comparative statics; e.g. Aiken and West 1991, p. 37; Dowling 2000,

pp. 284–291; Finney et al. 1984; Henseler and Fassott 2010, p. 728; Hirschey 2009, p. 99). Hence, the

slope of the line in Fig. 2.4 (i.e. @P @D= ¼ d þ m �M) is equal to the direct effect of discretion on

performance (d)when themoderator variable is zero (M ¼ 0) and is adjusted upwards or downwards to

the extent that themoderator diverges from zero.Moderating effects (m) therefore adjust the impact of


managers use discretion to increase, not alter or decrease performance, moderating

effects can potentially help explain the contradictory findings of the discretion puzzle(see Sects. 1.1 and 2.2.2), i.e. why existing studies have found evidence that

discretion increases, does not alter, and decreases performance. For this reason,

moderating effects are derived below from principal-agent theory and these

moderating effects are synthesised into the present study’s Hypotheses 2 and 3 in

Sect. 2.4.2, which work towards resolving the discretion puzzle.

The question then arises what moderator variables (M; or influences on

managers) influence whether managers use discretion in a more positive or negative

way. This section derives a number of such potential moderators from principal-

agent theory, while recognising that there might be other moderators and that the

nature of the moderating effects might differ across alternative specifications of

principal-agent models. Given that the negative predicted impact of discretion on

performance was shown above to result from the theory’s four assumptions, the

extent to which this impact is negative can be influenced by the extent to which

these assumptions apply. Consequently, moderating effects are derived from thefour assumptions of principal-agent theory, as these affect the extent to which

discretion reduces performance. A detailed derivation of moderating effects is

presented in Box 2.2. The resulting effects are summarised verbally below and

entered into Table 2.9 as well as Fig. 2.4 above (e.g. Caza 2007; 2011; Chang and

Wong 2003; Cheng et al. 2006; Davis et al. 1997b; Eisenhardt 1989; Fama

and Jensen 1983b; Finkelstein and Peteraf 2007; Jensen and Meckling 1976;

Xiaoyang Li 2007; Wang et al. 2008; Zhang and Li 2008b):

• From assumptions 1 and 4 regarding human behaviour (i.e. regarding the agent’s

utility function and interest in maximising performance relative to the principal),

the agent’s natural predisposition to performance maximisation as well as

compensation control mechanisms are derived as moderators for the impact of

discretion on performance in Box 2.2.66

• From assumption 2 concerning the principal’s asymmetric information on the

agent’s actions due to imperfect monitoring, the nature of managerial activities(i.e. complexity, uncertainty, and lack of observability) and monitoring controlmechanisms are found to potentially moderate discretion’s impact on perfor-

mance in Box 2.2. Whether these moderating effects are positive or negative is

shown to potentially depend on the above influences on managers related to the

agent’s utility function and interest in maximising performance (i.e. the agent’s

natural predisposition to performance maximisation and compensation control

mechanisms).

discretion onperformance upwards or downwards beyond the direct effect (d) and thus pivot the line inFig. 2.4, as indicated by the curved arrow therein.66 For example, when the plant manager’s relative interest in performance-maximisation (M )

increased (whether due to natural predisposition or compensation control mechanisms), the impact

of discretion on performance (@P @D= ¼ d þ m �M ) could become less negative, because the

plant manager would then use discretion to produce outcomes that were to a lesser extent harmful

to organisational performance.


• From assumption 3, which holds that the agent’s managerial action tends to

produce the outcomes intended, the agent’s ability, knowledge, and informationare likewise derived as potential moderators in Box 2.2. Once again, whether

these moderating effects are positive or negative can potentially depend on the

above influences on managers regarding human behaviour (i.e. the agent’s

natural predisposition to performance maximisation and compensation control

mechanisms).

This discussion of moderating effects in principal-agent theory is revisited in

Sect. 2.4.2 so as to derive the present study’s Hypothesis 2 and Hypothesis 3,67

whereas the discussion of direct effects in principal-agent theory contributes to

Hypothesis 4 and the theory’s unidimensional treatment of discretion contributes to

Hypothesis 1.

Box 2.2: Potential Moderating Effects Derived from Principal-Agent Theory

This Box takes the four aforementioned assumptions of principal-agent

theory in turn so as to derive a number of potential moderators from

principal-agent theory. These moderators (or influences on managers) are

summarised in Table 2.9 and contribute to the formulation of the present

study’s Hypotheses 2 and 3 (see Sect. 2.4.2). While the derivation is

conducted using the present study’s unit of analysis as an example (i.e. the

plant manager in China; see Chap. 3), it is recognised that there might be

other moderators and that the nature of the moderating effects might differ

across alternative specifications of principal-agent models.

• Assumption 1 and assumption 4 regarding the principal’s and the agent’sutility functions and interests state that both the principal and the agent seekto individually maximise their self-serving utility functions (assumption 1;

e.g. Albanese et al. 1997, p. 609; Chang and Wong 2003, p. 25; Davis et al.

1997b, pp. 20–22; Dicke and Ott 2002, p. 464; Eisenhardt 1989, pp. 58–59;

Jensen and Meckling 1976; Jensen and Murphy 1990, pp. 225–226;

Levinthal 1988, p. 153; Spremann 1987, p. 3; Thomsen and Pedersen 2000,

p. 690; Van Slyke 2007, p. 162;Werner and Tosi 1995, p. 1673) and that the

principal’s interests are more aligned with maximising performance than

those of the agent (assumption 4; e.g. Chang and Wong 2003; Eisenhardt

1989; Jensen 1986; Jensen and Meckling 1976; Jensen and Murphy 1990;

Shleifer and Vishny 1997; Spremann 1987; Thomsen and Pedersen 2000).(continued)

67 Hypothesis 2 and Hypothesis 3 regarding the respective moderating effects of firm type and firm

size on the impact of discretion on performance (see Sect. 2.4.2) are methodologically motivated

by the discussion of moderating effects in Box 2.2. In particular, principal-agent theory was shown

to imply that one moderator variable (i.e. the agent’s interests) can moderate both the direct effect

of discretion on performance (i.e. d > 0 versus d < 0 ) and the moderating effects of other

moderator variables (i.e. m > 0 versus m < 0). In parallel, the present study specifies that one

moderator variable (i.e. firm type) can moderate both the direct effect of discretion on performance

and the moderating effect of firm size (see Sect. 2.4.2).


As granting the agent additional discretion was shown above to shift

outcomes closer to those desired by the agent, the less performance-

maximising the interests of the agent are, the lower the performance may

be. The corollary is that relatively more performance-aligned interests of the

agent may mitigate the predicted negative impact of discretion on perfor-

mance. Hence, principal-agent theory implies that the agent’s interest inmaximising performance (relative to the principal; denoted by M ) may

have a positive moderating effect on the impact of discretion on performance

(i.e. m > 0).68 The agent’s interest in performance is in turn affected by the

agent’s natural predisposition to performance maximisation (which is fur-

ther addressed by stewardship theory in Sect. 2.3.3 below) and by the extent

to which the principal implements compensation control mechanisms thatprovide financial incentives and sanctions so as to align the interests of the

agent with maximising performance (see Sect. 2.3.2.1; e.g. Agrawal and

Knoeber 1996, p. 378; Berger et al. 1997, p. 1411; Burkart et al. 1997, p.

705; Chang andWong 2003, p. 6; Cheng et al. 2006; Eisenhardt 1989, p. 60;

Fama and Jensen 1983a, p. 345; Jensen and Meckling 1976; Jensen and

Murphy 1990, p. 226; Levinthal 1988, p. 153; Spremann 1987, p. 10;

Wang et al. 2008; Werner and Tosi 1995, p. 1673; Zhang and Li 2008b).

Eisenhardt (1989, p. 60) and Jensen and Meckling (1976) make this point,

positing that aligning the agent’s interests with those of the principal (e.g. by

outcome-based contracts or managerial shareholdings) tends to reduce

opportunism, i.e. makes it more likely that the agent behaves in the

principal’s interests. In the empirical literature (see Sect. 2.2), Caza (2007;

2011) has tested for the moderating effect of the agent’s natural predisposi-

tion to performance maximisation in terms of managerial commitment

(which was found to be insignificant) and Zhang and Li (2008b), Cheng

et al. (2006), andWang et al. (2008) have tested for the moderating effects of

compensation control mechanisms in terms of various managerial incentive

systems (which were found to be significant).

• Assumption 2 regarding the principal’s asymmetric information on theagent’s actions due to imperfect monitoring has been shown above to

constitute a source of discretion (i.e. antecedent)—whether due to the nature

ofmanagerial activities (i.e. complexity, uncertainty, and lack of observabil-

ity; Finkelstein and Peteraf 2007, pp. 237–243) or due to the monitoring

68 It was explained above that in the extreme case where the agent’s interest in performance

surpasses the principal’s interest in performance (i.e.M becomes large), the total impact of discretion

on performance might become positive. This is consistent with the described moderating effect,

since asM becomes large for a firm in a sample with a highly performance-aligned agent, the positive

moderating effect (m) in this firm becomes so large that it outweighs the overall negative direct effect

of discretion on performance (d < 0) estimated for the overall sample, i.e. @P @D= ¼ d þ m �M > 0

becomes positive for the performance-aligned agent despite d < 0.


control mechanisms in place (e.g. Berger et al. 1997, p. 1411; Burkart et al.1997, p. 705; Caza 2007; 2011; Chang and Wong 2003, p. 6; Eisenhardt

1989, p. 60; Fama and Jensen 1983b; Levinthal 1988, p. 153; Spremann

1987, pp. 10–11; Verhoest 2003, pp. 2–5;Wang et al. 2008;Werner and Tosi

1995, p. 1673): The less well the principal can monitor the agent’s actions,

the more discretion the agent may perceive. This is summarised in the

column on antecedents in Table 2.9. Yet in addition to affecting how much

managerial discretion the agent perceives (acting as an antecedent), theasymmetry of information may also influence how the agent uses his/her

discretion to impact on performance (acting as a moderator). These two

effects are exemplified below. Consider a plant manager in a firm with weak

monitoring control mechanisms (i.e. infrequent audits and performance

evaluations, denoted by a low value of M ¼ 0 ). If the plant manager’s

discretion for capital investments is increased from zero to 10,000 RMB,

then the plant manager might choose investments up to the value of 10,000

RMB that fulfil his/her self-serving interests (e.g. modern machinery)

rather than improve performance. With weak monitoring control

mechanisms (M ¼ 0), the impact of discretion on performance might then

be negative, i.e. @P @D= ¼ d þ m �M ¼ d þ m � 0 ¼ d < 0. When stricter

monitoring control mechanisms are implemented (denoted by a higher value

ofM ¼ 1), the aforementioned two effects might take place:

1. The monitoring control mechanisms could reduce the discretion per-

ceived by the plant manager (acting as an antecedent): Among capital

investments up to the value of 10,000 RMB, the plant manager’s multi-

ple choices are reduced to only such investments that ‘lie within the zone

of acceptance of powerful parties’ (principal, i.e. top management in

China; Hambrick and Finkelstein 1987, p. 378), which is narrowed due

to improved monitoring, i.e. less asymmetric information.

2. The monitoring control mechanisms could alter the way in which the plant

manager uses his/her discretion to affect performance (acting as a moder-ator). If the plant manager’s discretion for capital investments was

increased by another 10,000 RMB to 20,000 RMB, then with the stricter

monitoring control mechanisms in place (M ¼ 1) the plant manager could

not as freely choose investments out of self-serving interests at the cost of

performance as with weaker monitoring (M ¼ 0). Discretion would then

have a less negative impact on performance with better monitoring (M ¼ 1)

than weaker monitoring—so better monitoring would have a positive

moderating effect on the impact of discretion on performance (m > 0).69

(continued)

69 Algebraically, the impact of discretion on performance is @P @D= ¼ d þ m �M ¼ d þ m � 0 ¼ dwith weaker monitoring (M ¼ 0 ) and @P @D= ¼ d þ m �M ¼ d þ m � 1 ¼ d þ m with better

monitoring (M ¼ 1). The moderating effect on the impact of discretion on performance from

imposing better monitoring (i.e. from increasing M from M ¼ 0 to M ¼ 1) is thus equal to m (i.e.


The preceding explanation that under the assumptions of principal-

agent theory monitoring control mechanisms may have a negative effectas an antecedent on discretion and a positive effect as a moderator on the

impact of discretion on performance can be applied to the nature of

managerial activities (see above). As the nature of managerial activities

intensifies the asymmetry of information (i.e. due to higher complexity,

uncertainty, and lack of observability; Finkelstein and Peteraf 2007,

pp. 237–243), it has the opposite effects of improving monitoring control

mechanisms, which mitigate the asymmetry of information (e.g. Berger

et al. 1997, p. 1411; Burkart et al. 1997, p. 705; Caza 2007; 2011;

Chang and Wong 2003, p. 6; Eisenhardt 1989, p. 60; Fama and Jensen

1983b; Levinthal 1988, p. 153; Spremann 1987, pp. 10–11; Verhoest 2003,

pp. 2–5; Wang et al. 2008; Werner and Tosi 1995, p. 1673). The nature of

managerial activities may thus have a positive effect as an antecedent ondiscretion and a negative effect as a moderator on the impact of discretion

on performance—as noted in Table 2.9. This is also argued by Eisenhardt

(1989, p. 60) and Fama and Jensen (1983b), who contend that improving

the principal’s information to verify the agent’s behaviour makes it more

likely that the agent behaves in the interests of the principal. Empirically,

Caza (2007; 2011) and Wang et al. (2008) have tested for such moderating

effects of improved monitoring control mechanisms (in terms of the

number of similar units and performance monitoring, respectively),

which has yielded a mix of significant and insignificant results.

It should be noted that if instead of assumption 4, it is assumed that the

plant manager (agent) has more performance-maximising interests than top

management (principal), then these moderating effects are reversed to a

negative effect for monitoring control mechanisms (m < 0) and a positive

effect for the nature (i.e. complexity, uncertainty, and lack of observability)

of managerial activities (m > 0). The reason is that (as described in Box 2.1)

reducing the asymmetry of information brings outcomes closer to those

desired by the principal, which then entail lower performance. In this case,

reducing the asymmetry of information dampens the positive effect of discre-

tion on performance, since it leads the plant manager to choose actions that

are more in the interest of top management and therefore entail lower

performance.

• Assumption 3 that the agent’s managerial action is rational in terms oftending to produce the outcomes that the agent intends (e.g. Albanese et al.1997, p. 610; Davis et al. 1997b, pp. 20–22;Werner and Tosi 1995, p. 1673)

implies additional potential moderating effects on the impact of discretion

on performance. The agent’s ability, knowledge, and information to

the difference between these two equations). Hence, the impact of discretion will be less negative

with better monitoring to the extent that the moderating effect is positive (m > 0).


achieve his/her desired outcomes may moderate the extent to which the

agent can use his/her discretion to achieve desired outcomes. Specifically, if

as in principal-agent theory the agent’s interests are not well aligned with

performance maximisation (assumption 4), then improving the agent’sability, knowledge, and information (from M ¼ 0 to M ¼ 1) allows the

agent to better achieve his/her desired outcomes with a given level of

discretion, which could then ceteris paribus move outcomes further away

from performance. Then, the agent’s ability, knowledge, and information

could negatively moderate the impact of discretion on performance (m < 0).

The corollary is that when the agent is more performance-maximising than

the principal, so that the agent uses discretion to diligently work towards his/

her personal objectives and performance (Chang andWong 2003), the extent

to which performance is increased by discretion could be positively

moderated by the agent’s ability, knowledge, and information (m > 0).70

One might then argue that in the extreme case where agents are willing to

improve performance (d > 0) but a particular agent lacks the ability to do

so (M ¼ �1 ), the total impact of discretion on performance could be

negative (@P @D= ¼ d þ m �M ¼ d þ m � ð�1Þ� ¼ d � m). This is consis-tent with Davis et al.’s (1997b, pp. 23–24) observation that ‘[t]here are

many reasons other than poor motivation for agents’ failing to deliver high

performance for their principals (e.g., low ability, lack of knowledge, and

poor information).’ Likewise, Caza (2007; 2011) supports these postulated

effects by empirically testing for the moderating effects of the manager’s

experience and education. The findings confirm that managerial experience

has a significantly positive moderating effect (while managerial education(continued)

70 This three-way interaction (Henseler and Fassott 2010, p. 722) in the sense that the agent’s

interests influence whether ability, knowledge, and information improve or dampen the impact of

discretion on performance is consistent with seemingly opposing views in the literature. Hayek

(1945, pp. 524–526) advocates extending discretion downwards along the line of control (i.e.

decentralisation) on the premise that this may facilitate the use of everyone’s specialised knowl-

edge. On a similar note, Li (2007) argues that the board of directors (i.e. principal) may hire a

general manager (i.e. agent) as the manager possesses managerial expertise. Caza (2007, p. 13)

also contends that extending discretion may improve performance only if the manager’s ability

(e.g. education and experience) is sufficient so as to use discretion effectively for improving

performance. However, Chang and Wong (2003, p. 24) explain that managers ‘would not make

productive use of their expertise to improve firm performance if their self-interests were not

somehow tied to firm performance.’ In other words, only if a manager’s interests are sufficiently

aligned with improving performance may the manager’s ability, knowledge, and information

positively moderate the way managers use discretion to affect performance. If instead the

manager’s interests are poorly aligned with performance, then greater discretion that allows

managers to ‘respond more quickly to changing circumstances’ may be abused by the manager

so as to fulfil his/her private interests at the expense of those of the principal (Cheng et al. 2006,

pp. 341–342).


is insignificant), whereby performance is better only when managers are

simultaneously experienced and have more discretion.

These potential moderating effects (summarised in Table 2.9) enter into the

present study’s theoretical discretion model in Sect. 4.1 and are relevant for thederivation of Hypothesis 2 and Hypothesis 3 in Sect. 2.4.2. They are also

entered into the fictitious scatter plot in Fig. 2.4 above as influences that affect

the way that the manager uses his/her discretion to impact on performance and

therefore potentially pivot the downwards-sloping line. Based on the preceding

discussion, it can now be understood why—as claimed in Box 2.1—in excep-

tional cases principal-agent theory can be consistent with an insignificant

impact of discretion on performance. When a firm implements highly effective

compensation control mechanisms and monitoring control mechanisms, these

may help align the agent’s interest with performance maximisation (i.e. posi-

tive moderating effect from assumptions 1 and 4) and mitigate the principal’s

asymmetry of information (i.e. positive moderating effect from assumption 2).

The impact of discretion on performance (@P @D= ¼ d þ m �M) may then be

increased by the positive moderating effects (m) of the control mechanisms (M),

which would pivot the downwards-sloping line in Fig. 2.4 upwards. When

the control mechanisms are completely effective (i.e. M is high), then in the

limit the downwards-sloping line approaches the horizontal axis in Fig. 2.4 as

the impact of discretion on performance approaches zero (@P @D= ¼ d þ m�M � 0). In this exceptional case of complete controls, principal-agent theory

could then be consistent with an insignificant impact of discretion on perfor-

mance, as noted in Box 2.1 (e.g. Agrawal and Knoeber 1996, p. 377; Chang and

Wong 2003, p. 7).

Finally, this Box scrutinises potential interactions among the aforemen-

tioned moderating effects, which are built into the study’s theoretical discre-

tion model (see Sect. 4.1) and methodologically motivate Hypotheses 2 and 3.

The complexity inherent in this discussion on the impact of managerial

discretion on performance is that the agent’s interest in maximising perfor-

mance may affect both the direction of the direct effect of discretion on

performance (i.e. d > 0 versus d < 0) and the direction of moderating effects

(i.e. m > 0 versus m < 0) of the nature of managerial activities, monitoring

control mechanisms, and the agent’s ability, knowledge, and information.

For instance, when the agent’s interests are more aligned with performance

maximisation than those of the principal, it was explained above that discre-

tion is predicted to increase performance (d > 0) and that this increase in

performance might then be stronger, the greater the agent’s ability is (m > 0)

(e.g. Caza 2007; 2011; Chang and Wong 2003; Xiaoyang Li 2007). By

contrast, when the agent is less performance-maximising than the principal

(as assumed in most of principal-agent theory; Thomsen and Pedersen 2000,

p. 690), then discretion is predicted to decrease performance (d < 0), which is

intensified by the agent’s ability (m < 0) (e.g. Chang and Wong 2003, p. 24;


Cheng et al. 2006, pp. 341–342; Davis et al. 1997b, pp. 23–24). Methodolog-

ically, this is a case of three-way interaction (Henseler and Fassott 2010,

p. 722), i.e. the agent’s interests moderate (not only the direct effect d of

discretion on performance but also) the moderating effect (m ) of e.g. the

agent’s ability on the impact of discretion on performance (see Sect. 4.3).

This would be a three-way interaction between the agent’s interests, the

agent’s ability, and his/her discretion. The way that such potential three-

way interactions are treated in the present study is to combine the multi-

group comparison approach for testing moderating effects (e.g. Arnold 1982;

Henseler and Fassott 2010, pp. 719–721; Rigdon et al. 1998, p. 1;

Venkatraman 1989, p. 426) with the product term approach for testing

moderating effects (e.g. Carte and Russell 2003, pp. 480–495; Chin et al.

2003, pp. 196–200; Irwin and McClelland 2001, p. 105), as described for the

present study’s model in Chap. 4. Applying this methodology, the potential

predictions derived above from principal-agent theory for the impact of

discretion on performance can be synthesised as follows:71

• If the agent’s interest in maximising performance is less than the principal’s(due to the agent’s natural predisposition to performancemaximisation and/or

compensation control mechanisms in Table 2.9; denoted by superscript I),then principal-agent theory predicts a negative direct effect of discretion on

performance (dI < 0), a negative moderating effect of the nature of manage-

rial activities (mIact: < 0), a positive moderating effect of monitoring control

mechanisms (mImon: > 0), and a negative moderating effect of the agent’s

ability, knowledge, and information (mIabi: < 0; see above). Based on Equa-

tion (2.2) in Sect. 2.2.1 (extended to multiple moderators), the impact of

discretion on performance then becomes:

@P

@D

��

I

¼ dI|{z}

<0

þ mIact:

|ffl{zffl}

<0

� MIact: þ mI

mon:|ffl{zffl}

>0

� MImon: þ mI

abi:|ffl{zffl}

<0

� MIabi: < 0 (2.3)

which is the case commonly assumed in principal-agent theory and displayed

in Fig. 2.4 and Table 2.9.

• If the agent’s interest in maximising performance is greater than theprincipal’s (denoted by superscript II), then principal-agent theory would

predict a positive direct effect of discretion on performance (dII > 0), a

positive moderating effect of the nature of managerial activities (mIIact: > 0),

(continued)

71 It should be noted that the moderating effect of the agent’s interests could also be modelled as a

continuous (rather than a dichotomous) moderator variable. However, the dichotomy of the multi-

group comparison approach is chosen instead in order to ensure continuity with the present study’s

new model in Chap. 4.


a negative moderating effect of monitoring control mechanisms (mIImon: < 0),

and a positive moderating effect of the agent’s ability, knowledge, and

information (mIIabi: > 0; see above). Based on Equation (2.2) in Sect. 2.2.1

(extended to multiple moderators), the impact of discretion on performance

becomes:

@P

@D

��

II

¼ dII|{z}

>0

þ mIIact:

|ffl{zffl}

>0

� MIIact: þ mII

mon:|ffl{zffl}

<0

� MIImon: þ mII

abi:|ffl{zffl}

>0

� MIIabi: > 0 (2.4)

• The moderating effect of the agent’s interest in maximising performance isthen given by the difference in the parameters between the groups I and II(e.g. Arnold 1982; Henseler and Fassott 2010, pp. 719–721; Rigdon et al.

1998, p. 1; Venkatraman 1989, p. 426). For example, the moderating effect

of the agent’s interest being more (i.e. group II) rather than less (i.e. group I)aligned with performance maximisation on the total impact of discretion on

performance (@P @D= ) is the difference between Equation (2.4) and Equa-

tion (2.3), i.e. @P @D= jII � @P @D= jI > 0 . The moderating effect of the

agent’s interests on the impact of discretion on performance is thus

modelled to be positive, just as qualitatively described above. Likewise,

the moderating effect of the agent’s interests on the moderating effect of the

agent’s ability on the impact of discretion on performance (i.e. the three-way

interaction) is then predicted to be positive, i.e. mIIabi: � mI

abi: > 0, which

again matches the qualitative predictions that were discussed above.

In short, principal-agent theory was shown to imply that one moderator

variable (i.e. the agent’s interests) can moderate both the direct effect of

discretion on performance (i.e.d > 0versusd < 0) and the moderating effects

of other moderator variables (i.e. m > 0 versus m < 0). This motivates the

present study’s new model (see Chap. 4) to specify that one moderator

variable (i.e. firm type from Hypothesis 2) moderates both the direct effect

of discretion on performance and the moderating effect of firm size (from

Hypothesis 3), as Sect. 2.4.2 describes.

In sum, this section has reviewed the literature on principal-agent theory briefly in

general terms (e.g. historical context, unit of analysis, assumptions, and predictions)

as well as in detail with respect to the theory’s predicted impact of discretion on

performance. As shown in the literature review summary Table 2.10, principal-agent

theory is found to be applicable to the present study’s unit of analysis, namelymiddle

management in China.Moreover, although principal-agent theory contributes less to

the theoretical context of discretion (i.e. the construct and antecedents of discretion)

than managerial discretion theory, it postulates theoretical mechanisms for a mostly

negative impact of discretion on performance (i.e. for the consequences of


discretion). Specifically, the theory’s assumptions and postulated mechanisms have

allowed this section to derive a number of potential moderating effects implicit in

principal-agent theory that were summarised in Table 2.9 above. While fully testing

these moderating effects is not part of the research objective, the detailed discussion

has important implications for formulating the present study’s hypotheses (see

Sect. 2.4) as well as for developing the present study’s new model in Chap. 4.

Finally, the theoretical mechanisms derived from principal-agent theory form the

basis for discussing the predictions of stewardship theory in the next section (see

Sect. 2.3.3).

Table 2.9 Potential moderating effects derived from principal-agent theory

Assumptions Influences on managers

Effects as

antecedents

Effects as

moderators

Assumptions 1 and 4 on the

agent’s utility function and

interest in maximising

performance relative to the

principal (i.e. human

behaviour)

• The agent’s natural

predisposition to

performance

maximisation

• Positive

• Compensation control

mechanisms

• Positive

Assumption 2 on the

principal’s asymmetric

information regarding the

agent’s actions due to

imperfect monitoring

• The nature of

managerial activities

(complexity,

uncertainty, lack of

observability)

• Positive • Negative

(assumption 4)a

• Monitoring control

mechanisms

• Negative • Positive

(assumption 4)a

Assumption 3 on the agent’s

managerial actions tending

to produce his/her intended

outcomes

• The agent’s ability,

knowledge, and

information

(Added based on

managerial discretion

theory in Fig. 4.4 in

Sect. 4.1.2)

• Negative

(assumption 4)a

Source: Principal-agent theory including Zhang and Li (2008b), Caza (2007; 2011), Chang and

Wong (2003), Cheng et al. (2006), Davis et al. (1997b), Eisenhardt (1989), Fama and Jensen

(1983b), Finkelstein and Peteraf (2007), Jensen and Meckling (1976), Wang et al. (2008), and

Xiaoyang Li (2007)

Note: Table 2.9 summarises the potential moderating effects derived from principal-agent theory

in Box 2.2. It is recognised that there might be other moderators and that the nature of the

moderating effects might differ across alternative specifications of principal-agent models. The

column ‘Effects as Antecedents’ shows that the nature of managerial activities and monitoring

control mechanisms may respectively have positive and negative impacts on the level of perceived

managerial discretion. The column ‘Effects as Moderators’ shows that the listed influences may

also have positive and negative moderating effects on the impact of discretion (D) on performance

(P), i.e. on @P @D= (see Equation (2.2) in Sect. 2.2.1)a These signs apply under assumption 4. However, if one instead assumes that the agent is more

performance-maximising than the principal, then, as explained in Box 2.2, these moderating

effects may be reversed to a negative moderating effect of monitoring control mechanisms and

positive moderating effects of the nature of managerial activities and the agent’s ability, knowl-

edge, and information. There may thus be three-way interactions between these moderators and

the moderators related to assumptions 1 to 4 in Table 2.9 (see Box 2.2)



Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality







Impact on

performance:

• Direct

• Moderating

Contradictory

evidence

• Positive,

neutral,

negative

• Moderators

potentially

important

but rarely

modelled

(esp. firm

type and

firm size)

Theoretical

context specified

• Not specified

• Not specified.

Antecedents:

environmental,

organisational,

and managerial

Mechanisms

specified

• Negativea


management,

rarely China

Extended from

top to middle

management

Extended from

top to middle

management

Section 2.3.3

Section 2.4

Hypothesis 1

Middle

management

in China

Hypothesis 2

Hypothesis 3

Hypothesis 4

interests

relative to

principal’s

and asymmetry

of information

• E.g. agent’s

a As explained in Sect. 2.3.2, it is possible to assume that the principal is less performance-




generally contends that managerial discretion has a negative direct effect on performance (e.g.

Caza 2007, p. 10; Caza 2011; Chang and Wong 2003, p. 7; Davis et al. 1997b, p. 38;

Hutzschenreuter and Kleindienst 2007, p. 4; Jensen and Murphy 1990; Zhao et al. 2010)


2.3.3 Stewardship Theory

In parallel to the preceding sections on managerial discretion theory and principal-

agent theory, this section reviews the literature on stewardship theory in two parts:

• Section 2.3.3.1 explains stewardship theory briefly in terms of its historical

context, unit of analysis, assumptions, and predictions.

• Section 2.3.3.2 explores stewardship theory with respect to explaining the

impact of managerial discretion on performance. As in the previous sections,

this includes a discussion of the construct of discretion (i.e. definition and

dimensionality), the antecedents of discretion, and the consequences of discre-

tion (i.e. direct effects and moderating effects).

2.3.3.1 Description of TheoryAs to its historical context, stewardship theory evolved as an alternative to

principal-agent theory (see Sect. 2.3.2) for analysing situations in which a principal

(e.g. an owner) has a contract that delegates responsibility to a steward (e.g. a

manager) to act on the principal’s behalf and in which—in contrast to principal-

agent theory—the steward is motivated to act in the best interest of the principal

(e.g. Albanese et al. 1997; Arthurs and Busenitz 2003; Corbetta and Salvato 2004;

Davis et al. 1997a, b; Dicke and Ott 2002; Donaldson 1990; Donaldson and Davis

1989, 1991, 1993, 1994; Eddleston and Kellermanns 2007; Fox and Hamilton 1994;

Lane et al. 1999; Liu and Cai 2004; Miller and Le Breton-Miller 2006; Mills and

Keast 2009; Muth and Donaldson 1998; Salvato 2002; Tian and Lau 2001;

Tosi et al. 2003; Van Slyke 2007; Vargas Sanchez 2001, 2004, 2005;

Zahra 2003). Scholars in psychology and sociology have developed this alternative

theory in an effort to overcome the theoretical limitations of principal-agent theory

in terms of its economic assumptions regarding individualistic self-serving utility

maximisation (see assumption 1 and assumption 4 in Sect. 2.3.2), which have been

criticised for ignoring the complexities of organisational life and not holding true

for all managers (e.g. Arthurs and Busenitz 2003; Doucouliagos 1994; Frank 1994;

Graf Lambsdorff et al. 2006; Hirsch et al. 1987; Jensen and Meckling 1994;

Perrow 1986).

Concerning the unit of analysis, the initial primary focus of stewardship theory

was similar to that of principal-agent theory, namely on upper level managers (i.e.

top management) as stewards and owners as principals (see Davis et al. 1997b,

p. 21). Nevertheless, stewardship theory has subsequently been applied to a broad

range of contractual relations between shareholders, top management, middle

management, and employees in both the private and public sector, including within

family firms, between venture capitalists and entrepreneurs, and between public

managers and non-profit executive directors (e.g. Arthurs and Busenitz 2003;

Corbetta and Salvato 2004; Dicke and Ott 2002; Donaldson and Davis 1989,

1991, 1994; Eddleston and Kellermanns 2007; Fox and Hamilton 1994; Khanchel

2009; Miller and Le Breton-Miller 2006; Mills and Keast 2009; Muth and

Donaldson 1998; Salvato 2002; Tian and Lau 2001; Tosi et al. 2003; Van Slyke

2007; Vargas Sanchez 2001, 2004; Zahra 2003). In particular, stewardship theory is

applicable to stewards at the middle management level, as it applies to subordinates


of ‘principals who are managerial superordinates’ (Davis et al. 1997b, p. 25). It

follows that stewardship theory can be utilised for the unit of analysis of the present

study (i.e. the plant manager in China; see Chap. 3) in a similar way as can

managerial discretion theory and principal-agent theory (see Sects. 2.3.1 and

2.3.2). Although stewardship theory has overall received less attention in the

literature than the dominant principal-agent theory, it has been suggested that

stewardship theory is re-emerging in terms of popularity (Davis et al. 1997b,

p. 20; Dicke and Ott 2002, p. 464; Donaldson and Davis 1994, p. 151; Mills and

Keast 2009, pp. 10–13; Muth and Donaldson 1998, p. 5; Shleifer and Vishny 1997,

pp. 740–748; Thomsen and Pedersen 2000, p. 690).

Stewardship theory principally differs from principal-agent theory in its

assumptions regarding human behaviour (see assumption 1 and assumption 4 in

Sect. 2.3.2).72 Principal-agent theory from new institutional economics views

subordinates (i.e. agents) as individualistic, self-serving, and opportunistic,

whereas stewardship theory from psychology and sociology views subordinates

(i.e. stewards) as collectivists, pro-organisational, and trustworthy (e.g. Davis et al.

1997b, p. 20; Khanchel 2009, p. 98; Van Slyke 2007, p. 164; Vargas Sanchez 2004,

p. 3). The assumptions specifically differ as follows:

• Assumption 1. As to the first assumption in Sect. 2.3.2, principal-agent theory

assumes that the principal and the agent individually maximise their self-servingutility functions, which to the extent that these diverge leads the agent to

opportunistically pursue his/her self-serving interests at the expense of the

principal when given the opportunity to do so, thereby introducing a conflict

of interests (e.g. Albanese et al. 1997, p. 609; Chang and Wong 2003, p. 25;

Davis et al. 1997b, pp. 20–22; Dicke and Ott 2002, p. 464; Eisenhardt 1989,

pp. 58–59; Jensen and Meckling 1976; Jensen and Murphy 1990, pp. 225–226;

Levinthal 1988, p. 153; Spremann 1987, p. 3; Thomsen and Pedersen 2000,

p. 690; Van Slyke 2007, p. 162; Werner and Tosi 1995, p. 1673). By contrast,

while stewardship theory also assumes that stewards rationally maximise their

utility, the steward’s utility function is such that collectivistic, pro-organisational

behaviour yields higher utility than individualistic, self-serving behaviour,

and therefore stewards maximise their utility functions as they achieveorganisational instead of self-serving objectives (e.g. Davis et al. 1997b,

pp. 24–26; Eddleston and Kellermanns 2007, p. 549; Khanchel 2009,

pp. 98–99; Mills and Keast 2009, p. 13; Muth and Donaldson 1998, pp. 5–6).

Although a steward may have ‘survival’ needs, such as income, the steward thus

believes that his/her personal needs are met by working towards organisational

ends, which he/she associates with higher utility (e.g. Davis et al. 1997b, p. 25;

Dicke and Ott 2002, p. 464; Vargas Sanchez 2005, p. 19). In short, agents in

72Assumption 2 and assumption 3 from principal-agent theory are similar in stewardship theory,

namely that the principal cannot fully monitor the steward’s actions (i.e. asymmetric information)

and that the steward’s managerial action is rational in terms of tending to produce the outcomes

that the steward intends (e.g. Davis et al. 1997a, p. 612; Davis et al. 1997b, pp. 23–24; Fox and

Hamilton 1994, p. 78).


principal-agent theory are assumed to maximise their self-serving utility

functions, whereas stewards in stewardship theory are assumed to maximise

their pro-organisational utility functions.

• Assumption 4. As to the fourth assumption in Sect. 2.3.2, principal-agent theory

mostly assumes that the agent’s interests are less performance-maximising thanthe principal’s interests, meaning that the principal (e.g. owner) is more inter-

ested in the agent achieving high performance than the agent (e.g. manager)

himself/herself is (e.g. Chang and Wong 2003; Eisenhardt 1989; Jensen 1986;

Jensen and Meckling 1976; Jensen and Murphy 1990; Shleifer and Vishny 1997;

Spremann 1987; Thomsen and Pedersen 2000). By contrast, the steward’sinterests are at least as performance-maximising as the principal’s interests(e.g. Albanese et al. 1997, p. 609; Davis et al. 1997b, pp. 24–25; Dicke and Ott

2002, p. 464; Donaldson 1990, p. 377; Donaldson and Davis 1994, p. 159;

Khanchel 2009, p. 98; Liu and Cai 2004, pp. 3–4; Tosi et al. 2003, p. 2054;

Van Slyke 2007, p. 164; Vargas Sanchez 2005, pp. 17–18):

– In general, stewardship theory assumes that the principal’s interests are

strongly aligned with organisational performance, no matter whether the

principal is an owner or a managerial superordinate (e.g. top management,

as in the present study; Davis et al. 1997b, p. 25). As the steward is assumed to

aspire to organisational performance as well (e.g. ‘sales growth or profitabil-

ity’; Davis et al. 1997b, p. 24; Donaldson and Davis 1994, p. 159; Liu and Cai

2004, pp. 3–4; Tosi et al. 2003, p. 2054; Vargas Sanchez 2005, pp. 17–18), the

steward’s and the principal’s interests are generally both aligned with

maximising performance (Albanese et al. 1997, p. 609; Donaldson 1990,

p. 377; Van Slyke 2007, p. 164).

– Even when there are multiple principals with competing interests (some being

more aligned with maximising performance than others), the steward seeks to

maximise his/her utility by working towards maximising organisational per-

formance, which the steward believes to be in the best interest of the

principals (Davis et al. 1997b, p. 25; Donaldson 1990, p. 377). Hence, even

when the steward’s and principal’s interests are not fully aligned, a steward’s

behaviour will not depart from the organisation’s best interests, such as

achieving high organisational performance (Davis et al. 1997b, p. 24;

Dicke and Ott 2002, p. 464).

In short, principal-agent theory assumes that agents individually maximise their

self-serving utility functions which are mostly less performance-maximising than

those of the principal, whereas stewardship theory assumes that stewards maximise

their utility functions as they achieve organisational instead of self-serving

objectives and the steward’s interests are at least as performance-maximising as

those of the principal.73

73 It was shown in Box 2.1 in Sect. 2.3.2.2 that the nature of the impact of discretion on

performance may depend on whose interests are more performance-maximising (assumption 4),

which is further discussed below.


These changes in the assumptions regarding human behaviour (from individual-

istic, self-serving, and opportunistic to collectivist, pro-organisational, and trust-

worthy) drastically alter the predictions about the manager’s (i.e. agent’s or

steward’s) actions:

• Principal-agent theory predicts that agents (e.g. plant managers) tend to abuse

discretion so as to pursue their own self-serving interests at the cost of the

principal (e.g. top management in China; see Sect. 2.3.2; Chang and Wong

2003, p. 6; Davis et al. 1997b, pp. 22–23; Jensen and Meckling 1976; Levinthal

1988, p. 153; Spremann 1987; Thepot 2007). When granting an agent additional

managerial discretion, the agent is thus predicted to use this greater latitude of

managerial action to choose actions that produce outcomes which are closer to

maximising the agent’s own utility although they tend to be further away from

maximising the principal’s utility. These outcomes (preferred by the agent) tend

to entail lower performance, since the agent’s interests are mostly assumed to be

less performance-maximising than the principal’s interests (see assumption 4 in

Sect. 2.3.2; e.g. Chang and Wong 2003; Eisenhardt 1989; Jensen 1986;

Jensen and Meckling 1976; Jensen and Murphy 1990; Shleifer and Vishny

1997; Spremann 1987; Thomsen and Pedersen 2000). Therefore, principal-

agent theory predicts that increasing an agent’s discretion tends to reduce

performance, whereas decreasing an agent’s discretion tends to improve perfor-

mance. Managerial discretion is thus predicted to mostly have a negative impact

on performance (e.g. Caza 2007, p. 10; Caza 2011; Chang and Wong 2003, p. 7;

Davis et al. 1997b, p. 38; Hutzschenreuter and Kleindienst 2007, p. 4; Jensen and

Murphy 1990; Zhao et al. 2010), which is recorded as the principal-agent

hypothesis for Hypothesis 4 in Sect. 2.4.3.

• Stewardship theory, by contrast, predicts that stewards (e.g. plant managers) do

not tend to abuse discretion to pursue their own self-serving interests at the

expense of the principal but instead use discretion to work diligently towards the

organisation’s best interests, such as performance, since, as assumed above,

stewards maximise their utility functions by achieving organisational instead

of self-serving objectives (e.g. Davis et al. 1997b, pp. 24–26; Dicke and Ott

2002, p. 464; Eddleston and Kellermanns 2007, p. 549; Khanchel 2009,

pp. 98–99; Mills and Keast 2009, p. 13; Muth and Donaldson 1998, pp. 5–6;

Vargas Sanchez 2005, p. 19). Even when the steward’s and principal’s interests

are equally performance-maximising (a situation in which the self-serving utility

function assumed in principal-agent theory predicts that discretion may not

affect performance; see Box 2.1 in Sect. 2.3.2.2), granting a steward additional

discretion is predicted to improve performance in stewardship theory.74

74 For example, consider granting a plant manager additional discretion for making capital

investments, i.e. increasing the maximum capital investment that the plant manager (i.e. agent

or steward) can undertake without prior authorisation from top management at corporate head-

quarters in China (i.e. principal) from 10,000 RMB to 20,000 RMB. With the self-serving utility

function assumed in principal-agent theory (see the first assumption above), if the plant manager’s

and top management’s interests were equally performance-maximising (see the fourth assumption


The reason is that the pro-organisational utility function assumed in stewardship

theory (assumption 1) implies an additional theoretical mechanism whereby

discretion improves performance by strengthening the steward’s motivation(regardless of whether the steward is equally or more performance-maximising

than the principal; see assumption 4 above).75 This additional theoretical mech-

anism derives from the assumption that stewards with their pro-organisational

utility functions (as opposed to agents with their self-serving utility functions)

are not motivated by individual goals but rather by intrinsic rewards, such as

discretion, trust, and job satisfaction (Corbetta and Salvato 2004; Davis et al.

1997b, p. 21; Donaldson and Davis 1991, p. 51, 1993, pp. 215–216; Eddleston

and Kellermanns 2007, p. 548; Khanchel 2009, p. 98; Liu and Cai 2004, pp. 3–4;

Tosi et al. 2003, p. 2054; Van Slyke 2007, pp. 164–165; Vargas Sanchez 2004,

p. 3). Implementing controls that reduce the steward’s discretion can thus

potentially be counter-productive by lowering the steward’s motivation

(Argyris 1964; Burkart et al. 1997, pp. 693–694; Davis et al. 1997b, p. 25;

Dicke and Ott 2002, p. 468; Frey 1993; Van Slyke 2007, pp. 162–163;

Vargas Sanchez 2005, p. 18). It follows that when granting a steward additional

discretion, the steward is predicted to use this greater latitude of action to

diligently choose actions that improve performance, which is generally assumed

to be in the principal’s best interest (Corbetta and Salvato 2004; Davis et al.

1997b, p. 25; Eddleston and Kellermanns 2007, p. 548). Hence, stewardship

theory predicts that increasing a steward’s discretion tends to improve perfor-

mance, whereas decreasing his/her discretion tends to reduce performance.

Managerial discretion is therefore predicted to have a positive impact on perfor-

mance, which is recorded for the present study’s Hypothesis 4 as the stewardship

hypothesis in Sect. 2.4.3.76

These opposing predictions of principal-agent theory and stewardship theory

(which derive from their differing assumptions regarding human behaviour, as

described above) lead to opposing prescriptions for principals in terms of how to

manage their managers (i.e. their agents and stewards):

• As the self-serving utility maximisation (assumption 1) in principal-agent theory

implies that agents tend to abuse discretion to pursue their own self-serving

above), then capital investments with the same performance consequences would be predictedwith and without the higher discretion (i.e. whether or not the plant manager needed to seek prior

authorisation from top management for the investments between 10,000 RMB to 20,000 RMB).

With the pro-organisational utility function assumed in stewardship theory, however, even if

interests were equally performance-maximising, then the higher discretion would strengthen theplant manager’s motivation (e.g. to make greater efforts to choose investments wisely) and thuswould be predicted to boost performance.75 This is the additional theoretical mechanism of stewardship theory which was mentioned in

Sect. 2.3.2.2 to help explain a positive impact of discretion on performance.76 Box 2.3 in Sect. 2.3.3.2 creates transparency on how two theoretical mechanisms translate the

assumptions of stewardship theory into this predicted positive impact of discretion on

performance.


interests at the cost of the principal (see above), principal-agent theorists recom-

mend that principals impose internal control mechanisms that constrain the

discretion and therefore the potential misconduct of the agents (see Sect. 2.3.2;

e.g. Agrawal and Knoeber 1996, p. 378; Berger et al. 1997, p. 1411;

Burkart et al. 1997, p. 705; Caza 2007; 2011; Chang and Wong 2003, p. 6;

Cheng et al. 2006; Eisenhardt 1989, p. 60; Fama and Jensen 1983a, p. 345,

Fama and Jensen 1983b; Jensen and Meckling 1976; Jensen and Murphy 1990,

p. 226; Levinthal 1988, p. 153; Spremann 1987, p. 10; Verhoest 2003, pp. 2–5;

Wang et al. 2008; Werner and Tosi 1995, p. 1673; Zhang and Li 2008b).

• As the pro-organisational utility maximisation (assumption 1) in stewardship

theory implies that stewards use discretion to work diligently towards the

organisation’s best interests to the benefit of the principal (see above), steward-

ship theorists prescribe that principals implement empowering governancemechanisms which enable the steward through discretion (e.g. Corbetta and

Salvato 2004; Davis et al. 1997b, pp. 25–26; Donaldson and Davis 1991, p. 52;

Eddleston and Kellermanns 2007, p. 547; Khanchel 2009, pp. 98–99; Liu and Cai

2004, p. 4; Mills and Keast 2009, pp. 14–15; Tosi et al. 2003, p. 2054; Van Slyke

2007, pp. 165–167; Vargas Sanchez 2004, p. 3, 2005, pp. 18–19).77 In steward-

ship theory, empowering governance mechanisms that increase the steward’s

discretion are preferable to control mechanisms that reduce the steward’s discre-

tion for two reasons. First, the steward’s interests are assumed in stewardship

theory to be aligned with the organisation’s and the principal’s interests (see

above),meaning that the costs to the principal of implementing compensation and

monitoring control mechanisms so as to guarantee pro-organisational behaviour

are unnecessary (e.g. Davis et al. 1997b, pp. 24–26; Mills and Keast 2009, p. 13).

Second, not only are control mechanisms unnecessarily costly, they are also

potentially counter-productive for stewards, since as described above they can

lower the steward’s motivation (Argyris 1964; Burkart et al. 1997, pp. 693–694;

Davis et al. 1997b, p. 25; Dicke and Ott 2002, p. 468; Frey 1993; Van Slyke 2007,

pp. 162–163; Vargas Sanchez 2005, p. 18). As stewards are motivated by intrinsic

rewards such as discretion and can be trusted, stewardship theory recommends

extending a steward’s discretion so as to maximise the benefits of a steward and

improve organisational performance (Corbetta and Salvato 2004; Donaldson and

Davis 1991, p. 51, 1993, pp. 215–216; Eddleston and Kellermanns 2007, p. 548;

Khanchel 2009, p. 98; Liu and Cai 2004, pp. 3–4; Tosi et al. 2003, p. 2054;

Vargas Sanchez 2004, p. 3).

Applied to the present study’s unit of analysis, principal-agent theory therefore

recommends top management at corporate headquarters in China (i.e. the principal)

to constrain the discretion of the plant manager so as to prevent potential miscon-

duct, whereas stewardship theory recommends top management to empower the

77 For instance, Donaldson and Davis (1991) recommend having a chief executive officer (CEO)

who is a steward as the chair of the board of directors, since this empowering governance

mechanism grants the CEO greater discretion to shape strategy in the company’s best interest

without fear of interference by an outside chair.


plant manager by granting him/her greater discretion so as to motivate and enable

him/her to strengthen the organisation’s performance. In order to create a better

understanding of how these conflicting recommendations might hold true, the next

section examines the literature on stewardship theory in greater detail with respect

to its explanation of the impact of managerial discretion on performance.

2.3.3.2 Analysis of Impact of Discretion on PerformanceScrutinising stewardship theory with respect to the impact of managerial discretion

on performance reveals that much like principal-agent theory (see Sect. 2.3.2) and

in contrast to managerial discretion theory (see Sect. 2.3.1), stewardship theory

tends not to focus on specifying the theoretical context of managerial discretion in

terms of defining the construct and antecedents of discretion, but rather focuses on

specifying discretion’s consequences by postulating mechanisms why the impact of

discretion on performance could be positive (rather than negative in principal-agent

theory) and how it might be moderated. These points generate implications for the

present study’s hypotheses in their attempt to work towards resolving the discretion

puzzle and are discussed in the same sequence as in the previous sections:



3. Consequences of managerial discretion (i.e. direct and moderating effects).

First, in line with principal-agent theory (see Sect. 2.3.2) and in contrast to

managerial discretion theory (see Sect. 2.3.1), stewardship theory has paid little

attention to specifying the construct of managerial discretion, as noted in Sect. 2.1.

Although stewardship theorists tend not to explicitly define discretion, they employ

the term in a way consistent with the definition of discretion in managerial discre-

tion theory (see Sect. 2.1.1; e.g. Davis et al. 1997b, pp. 25–26; Fox and Hamilton

1994, pp. 70–74; Hambrick and Finkelstein 1987, pp. 371–378; Mills and Keast

2009, pp. 14–15; Van Slyke 2007, pp. 165–167; Vargas Sanchez 2005, p. 19).

Moreover, concerning the dimensionality of discretion, stewardship theory has

treated discretion as a single (unidimensional) construct with a single positive

impact on performance (e.g. Dicke and Ott 2002, p. 468; Fox and Hamilton 1994,

p. 74; Vargas Sanchez 2005, p. 19). This is similar to the unidimensionality implicit

in principal-agent theory (e.g. Chang and Wong 2003; He et al. 2009;

Spremann 1987, p. 18; Xu et al. 2005) and different from the multidimensionality

in managerial discretion theory (e.g. Carpenter and Golden 1997, p. 195;

Caza 2007, pp. 26–82; Chen 2006; Finkelstein and Peteraf 2007, p. 245; Hambrick

and Abrahamson 1995, p. 1439; Hambrick and Finkelstein 1987, pp. 371–402;

Hambrick et al. 1993, p. 409; see Sect. 2.1.2). As explained in Sect. 2.1.2, this

discontinuity between the unidimensionality and multidimensionality of manage-

rial discretion motivates this study to test Hypothesis 1 (dimensions of discretion) in

an attempt to help resolve the discretion puzzle.

Second, antecedentsofmanagerial discretion (i.e. causes, determinants or sources of

discretion) tend not to be addressed in stewardship theory either—neither explicitly as

inmanagerial discretion theory (see Sect. 2.3.1; e.g. Acemoglu et al. 2007; Bloom et al.

2008; Carpenter andGolden 1997; Caza 2007; Cennamo et al. 2009; Cheng et al. 2006;

Colombo and Delmastro 2004; Crossland 2007; Crossland and Hambrick 2007;


Finkelstein and Boyd 1998; Hambrick and Abrahamson 1995; Hambrick and

Finkelstein 1987; Xiaoyang Li 2007; Yougen Li and Zhao 2004; Walters 1995;

Zhang et al. 2006a) nor implicitly as in principal-agent theory (see Sect. 2.3.2; e.g.

Eisenhardt 1989; Finkelstein and Peteraf 2007; Holmstrom 1979; Jensen and Murphy

1990; Rumelt 1984; Spremann 1987; Tirole 1988). As explained below, stewardship

theorists do employ the term antecedents for a number of psychological characteristics,

perceptions of the organisation’s situational characteristics, and expectations (e.g.

Davis et al. 1997b, pp. 27–43), yet these refer to ‘antecedents in the prediction of

stewardship versus agency relationships’ (Davis et al. 1997b, p. 37) rather than

antecedents in the prediction of the level of discretion. In other words, the antecedents

in stewardship theory do not refer to antecedents of managerial discretion but rather to

determinants of the nature of the consequences of managerial discretion (i.e.

determinants of whether the impact of discretion on performance is positive as in

stewardship relationships or negative as in agency relationships). Both of these types

of antecedents are integrated into the study’s new theoretical discretion model in

Sect. 4.1.2, namely as antecedents of discretion and as influences on managers with

moderating effects on the impact of discretion on performance.

Third, in contrast to managerial discretion theory (see Sect. 2.3.1), stewardship

theorists postulate mechanisms regarding the consequences of managerial discre-

tion that predict the impact of discretion on performance as being positive and

explain in which way this impact might be moderated. This prediction challenges

the prediction of principal-agent theory that the impact of discretion on perfor-

mance is mostly negative (e.g. Caza 2007, p. 10; Caza 2011; Chang and Wong

2003, p. 7; see Sect. 2.3.2; Davis et al. 1997b, p. 38; Hutzschenreuter and

Kleindienst 2007, p. 4; Jensen and Murphy 1990; Zhao et al. 2010). As described

in Sects. 2.2.1 and 2.3.2.2, the impact of discretion on performance can be

decomposed into the sum of the direct effect of discretion on performance (d )and the moderating effect of the moderator variable (m) multiplied by the value of

the moderator variable (M): @P @D= ¼ d þ m �M (see Equation (2.2) in Sect. 2.2.1

and Sect. 4.3.2 on comparative statics; e.g. Aiken and West 1991, p. 37;

Dowling 2000, pp. 284–291; Finney et al. 1984; Henseler and Fassott 2010,

p. 728; Hirschey 2009, p. 99). In line with the previous sections, the consequences

of discretion are described in terms of the overall direct effect of discretion on

performance before exploring moderating effects that may adjust the impact of

discretion above or below this direct effect.

Starting with the predicted direct effect of discretion on performance,

Sect. 2.3.3.1 has already explained how the differing assumptions regarding

human behaviour in stewardship theory and principal-agent theory lead to a

predicted positive effect (d > 0) and negative effect (d < 0) of discretion on

performance, respectively. This is summarised briefly using the example of the

plant manager in China (i.e. the unit of analysis of the present study; see Chap. 3)

and with reference to the fictitious scatter plot in Fig. 2.5, which helps take a step

towards defining the present study’s hypotheses and thereby towards resolving the

discretion puzzle. Figure 2.5 depicts a scatter plot similar to those in Figs. 2.1, 2.2,

2.3 and 2.4 in the previous sections. The upwards-sloping and downwards-sloping


lines represent performance (P) as a function of discretion (D) under stewardshiptheory (d > 0) and principal-agent theory (d < 0), respectively.78 The differences in

the predicted impact of discretion on performance according to principal-agent

theory and stewardship theory are then as follows:

• Principal-agent theory (see Sect. 2.3.2) predicts that granting the plant manager

additional discretion (moving rightwards on Fig. 2.5) will mostly reduce perfor-

mance (d < 0 ; the downwards-sloping line in Fig. 2.5), because the plant

manager individually maximises utility by pursuing his/her self-serving interests

(assumption 1) and given that top management at corporate headquarters in

China cannot fully monitor the plant manager’s actions (assumption 2), he/she

will choose to undertake such capital investments that produce the plant

manager’s desired outcomes (assumption 3) but tend to come at the expense of

the principal and reduce performance (assumption 4).

• Stewardship theory (e.g. Corbetta and Salvato 2004; Davis et al. 1997b,

pp. 25–26; Donaldson and Davis 1991, p. 52; Eddleston and Kellermanns

2007, p. 547; Khanchel 2009, pp. 98–99; Liu and Cai 2004, p. 4; Mills and

Keast 2009, pp. 14–15; Tosi et al. 2003, p. 2054; Van Slyke 2007, pp. 165–167;



HighManagerial

Discretion (D)

Low Performance (P)

Psychological CharacteristicsSituational CharacteristicsExpectations (Risk and Trust)


Source: Principal-agent theory and stewardship theory

78 The slopes of the lines measure the impact of discretion on performance (i.e.@P @D= ¼ d þ m �M)

and are therefore equal to the direct effect of discretion on performance (d) when the moderator

variable is zero (M ¼ 0), and are adjusted upwards or downwards to the extent that the moderator

diverges from zero.


Vargas Sanchez 2004, p. 3, 2005, pp. 18–19) postulates that extending the plant

manager’s discretion (moving rightwards on Fig. 2.5) will improve performance

(d > 0; the upwards-sloping line in Fig. 2.5), because the plant manager

maximises utility by achieving organisational instead of self-serving objectives

(assumption 1), so although top management cannot fully monitor the plant

manager’s actions (assumption 2), he/she will diligently choose such actions that

produce the plant manager’s desired outcomes (assumption 3), namely enhanc-

ing performance (assumption 4). As described in Box 2.3, there are two theoret-

ical mechanisms at work in stewardship theory that both tend to contribute to the

positive effect of managerial discretion on performance. Transparency on these

mechanisms enables the ensuing discussion to work towards resolving the

discretion puzzle (see Sects. 1.1 and 2.2.2).

With transparency on how the differences in assumptions 1 and 4 between

principal-agent theory and stewardship theory translate (via the two mechanisms

in Box 2.3) into a negative and positive predicted impact of discretion on perfor-

mance, respectively, it is possible to take a step towards resolving the discretionpuzzle. Specifically, the empirical evidence reviewed in Sect. 2.2.2 finding

positive impacts of discretion on performance (d > 0) (e.g. Agarwal et al. 2009;

Barnabas and Mekoth 2010; Chang and Wong 2003; Gammelgaard et al. 2010;

Khanchel 2009) can now be interpreted as supporting stewardship theory, whereas

the evidence finding negative impacts (d < 0) (e.g. He et al. 2009; Heinecke 2011;

Stano 1976; Williamson 1963; Xu et al. 2005) can be viewed in favour of principal-

agent theory. However, the coexistence of ample empirical evidence finding posi-

tive, negative, and also insignificant impacts (d � 0) (e.g. Caza 2011; Groves et al.

1994; Yougen Li and Zhao 2004; Lopez-Navarro and Camison-Zornoza 2003;

Venaik 1999) indicates that neither stewardship theory nor principal-agent theory

universally applies.79 On the contrary, the conflicting evidence suggests that for the

managers analysed in some empirical studies (d > 0 ; termed ‘stewards’) the

assumptions 1 and 4 of stewardship theory appear to apply, while for the managers

analysed in other empirical studies (d < 0 ; termed ‘agents’) the assumptions 1 and

4 of principal-agent theory appear to apply. A reason why evidence on the impact of

discretion on performance has ranged from positive to neutral and even negative

might then be that certain studies have focused on stewards as the unit of analysis

where the assumptions of stewardship theory apply (d > 0), others on agents where

the assumptions of principal-agent theory apply (d < 0), and yet others on a mix of

agents and stewards in their sample (d � 0) where positive and negative effects

average out to an insignificant effect.80

79 As explained in Sect. 2.3.1, managerial discretion theory does not make specific predictions

regarding the nature of the impact of managerial discretion on performance (i.e. whether it is

positive, neutral or negative; e.g. Adams et al. 2005; Crossland and Hambrick 2007; Finkelstein

and Hambrick 1990; Hambrick and Finkelstein 1987; Misangyi 2002; Quigley and Hambrick

2009; Tang 2008; Zhao et al. 2010).80 This extends the attempted explanation derived from principal-agent theory in Box 2.1 in

Sect. 2.3.2.2.


Box 2.3: Theoretical Mechanisms for Impact of Discretion on Performance

This Box creates transparency on two theoretical mechanisms that translate

the distinct assumptions of principal-agent theory and stewardship theory into

the respectively negative and positive predicted impacts of discretion on

performance. This enables the ensuing discussion in this section to work

towards resolving the discretion puzzle.

1. As described above (see assumption 4), principal-agent theory mostly

assumes that the agent’s interests are less performance-maximising than

the principal’s interests (e.g. Chang and Wong 2003; Eisenhardt 1989;

Jensen 1986; Jensen and Meckling 1976; Jensen and Murphy 1990;

Shleifer and Vishny 1997; Spremann 1987; Thomsen and Pedersen

2000), whereas stewardship theory assumes that the steward’s interests

are equally or more performance-maximising than the principal’s interests(e.g. Albanese et al. 1997, p. 609; Davis et al. 1997b, pp. 24–25; Dicke and

Ott 2002, p. 464; Donaldson 1990, p. 377; Donaldson and Davis 1994,

p. 159; Khanchel 2009, p. 98; Liu and Cai 2004, pp. 3–4; Tosi et al. 2003,

p. 2054; Van Slyke 2007, p. 164; Vargas Sanchez 2005, pp. 17–18). It was

explained in Sect. 2.3.2 that with greater discretion, the plant manager can

shift outcomes closer to outcomes he/she desires, which in principal-agent

theory tend to be less performance-maximising and in stewardship theory

equally or more performance-maximising than the outcomes desired by

the principal (e.g. Albanese et al. 1997, p. 610; Chang and Wong 2003,

p. 6; Davis et al. 1997b, pp. 20–23; Jensen and Meckling 1976;

Levinthal 1988, p. 153; Spremann 1987; Thepot 2007; Werner and Tosi

1995, p. 1673). Through this first mechanism, granting a plant manager

greater discretion for making capital investments could result in lower

performance (d < 0) if his/her interests were less performance-maximising

than the principal’s, similar performance (d � 0) if his/her interests were

equally performance-maximising, and higher performance (d > 0) if his/

her interests were more performance-maximising (see Box 2.1 in

Sect. 2.3.2.2). Although this mechanism can thus explain why the

differences in assumption 4 imply a negative impact of discretion

on performance (d < 0) in principal-agent theory, this mechanism on its

own implies that stewardship theory predicts a positive impact (d > 0)

when the steward is more performance-maximising than the principal and

a neutral impact (d � 0) when the steward is as performance-maximising

as the principal.

2. A second theoretical mechanism (derived from assumption 1) ensures that

stewardship theory predicts a positive impact of discretion on performance

(d > 0) even when the steward is equally performance-maximising as the

principal. As explained above, the pro-organisational utility functions thatstewards seek to maximise imply that discretion strengthens the steward’s

motivation (regardless of whether the steward is equally or more performance-

maximising than the principal; i.e. regardless of assumption 4), because with(continued)


these utility functions stewards are motivated by intrinsic rewards, such as

discretion, trust, and job satisfaction, rather than individual goals (Corbetta

and Salvato 2004; Davis et al. 1997b, p. 21; Donaldson and Davis 1991, p.

51, 1993, pp. 215–216; Eddleston and Kellermanns 2007, p. 548;

Khanchel 2009, p. 98; Liu and Cai 2004, pp. 3–4; Tosi et al. 2003,

p. 2054; Van Slyke 2007, pp. 164–165; Vargas Sanchez 2004, p. 3).

Granting a steward additional discretion therefore strengthens his/her

motivation to improve performance (Corbetta and Salvato 2004;

Davis et al. 1997b, p. 25; Eddleston and Kellermanns 2007, p. 548),

whereas constraining the steward’s discretion can lower the steward’s

motivation and consequently, his performance (Argyris 1964; Burkart

et al. 1997, pp. 693–694; Davis et al. 1997b, p. 25; Dicke and Ott 2002,

p. 468; Frey 1993; Van Slyke 2007, pp. 162–163; Vargas Sanchez 2005,

p. 18).81 By this additional mechanism derived from assumption 1, stew-

ardship theory can universally predict that discretion increases perfor-

mance (d > 0) even when the agent is equally performance-maximising

as the principal—a case where the above mechanism derived from

principal-agent theory would on its own predict a neutral effect of

discretion.

Applying these theoretical mechanisms to the unit of analysis of the

present study, it can be seen that even when the plant manager (agent) is

equally performance-maximising as top management in China (principal;

assumption 4), stewardship theory still predicts that granting discretion

ameliorates performance via the second mechanism. By virtue of the first

mechanism, a plant manager would be predicted to choose equally

performance-maximising capital investments whether or not he/she has to

seek top management’s approval (i.e. with or without discretion), so granting

additional discretion would not tend to affect performance (d � 0). Yet with

the second mechanism under assumption 1 of stewardship theory, the plant

manager with a pro-organisational utility function would be predicted to be

motivated by greater discretion, which would entice him/her to expend

greater efforts in choosing e.g. investments and thereby improving perfor-

mance (d > 0). Stewardship theory therefore predicts the plant manager to

use additional discretion to diligently choose actions that improve perfor-

mance (d > 0), no matter whether his/her interests are equally or more

aligned with maximising performance than those of top management.

81 This reasoning resonates with Aghion and Tirole’s (1997) view shared by e.g. Xiaoyang Li

(2007) and Xu et al. (2005) that extending discretion downwards along the line of control

(decentralisation) can raise the agent’s initiative to acquire productive information and facilitate

his/her participation in the contractual relationship.


Moving towards resolving the discretion puzzle would then require a more

granular distinction between units of analysis (as already motivated empirically in

Sect. 2.2) according to whether the assumptions of principal-agent theory or steward-

ship theory are more applicable. As explained in Sect. 2.2.4, this can equivalently be

achieved by either separately estimating the impact of discretion on performance for

multiple groups of the unit of analysis that are segmented according to these

assumptions (e.g. Arnold 1982; Henseler and Fassott 2010, pp. 719–721; Rigdon

et al. 1998, p. 1; Venkatraman 1989, p. 426) or by modelling the impact of discretion

on performance with moderator variables (M ) (e.g. Carte and Russell 2003, pp.

480–495; Chin et al. 2003, pp. 196–200; Irwin and McClelland 2001, p. 105) that

allow a model to predict different impacts of discretion on performance (@P @D= ¼ dþ m �M) according to the different values of the moderator variables (i.e. according

to different assumptions). The remainder of this section therefore derives moderator

variables (M) from the assumptions of stewardship theory and principal-agent theory

so as to allow the present study’s new theoretical discretion model in Sect. 4.1 to

differentiate the impact of discretion on performance according to these divergent

assumptions. The moderators derived then lead to the two moderators in Sect. 2.4.2

(i.e. Hypothesis 2: firm type and Hypothesis 3: firm size), which are empirically

tested within this study’s empirical discretion model. By testing whether the impact

of discretion on performance (@P @D= ) can differ by these moderator variables (M),

this study therefore establishes whether managers under different influences (i.e.

different firm types and firm sizes) tend to use their discretion more like stewards

or more like agents—thus leading to positive impacts in some situations and negative

impacts in others—and thereby takes a step towards fulfilling the research objective

of working towards resolving the discretion puzzle.

Before deriving new moderator variables (i.e. influences on managers) from the

assumptions of stewardship theory, the moderators derived from the assumptions of

principal-agent theory in Sect. 2.3.2.2 are briefly revisited.82 These moderators

include the manager’s natural predisposition to performance maximisation and

compensation control mechanisms (derived from assumptions 1 and 4), the natureof managerial activities andmonitoring control mechanisms (derived from assump-

tion 2), and the manager’s ability, knowledge, and information (derived from

assumption 3; see Table 2.9 in Sect. 2.3.2.2; e.g. Caza 2007; 2011; Chang and

Wong 2003; Cheng et al. 2006; Davis et al. 1997b; Eisenhardt 1989; Fama and

Jensen 1983b; Finkelstein and Peteraf 2007; Jensen and Meckling 1976;

Xiaoyang Li 2007; Wang et al. 2008; Zhang and Li 2008b).

82 As stated earlier,moderator variables (M)withmoderating effects (m)may influence the extent towhich

the impact of discretion on performance (@P @D= ¼ d þ m �M) is positive or negative, or equivalently,

the slopes of the lines in Fig. 2.5 (@P @D= ¼ d þ m �M). Strong moderating effects may therefore help

explain whether discretion increases ( @P @D= > 0 ) or decreases performance ( @P @D= < 0 ), or

equivalently, whether the line in Fig. 2.5 is upwards-sloping or downwards-sloping. Moderating effects

might thus potentially explain in which cases (i.e. for which values ofM) the predictions of stewardship

theory versus principal-agent theory hold true, thus paving away for reconciling the theories and resolving

the discretion puzzle.


Among these potential moderator variables, stewardship theorists have

concentrated on themanager’s natural predisposition to performance maximisation(derived from assumptions 1 and 4 on human behaviour in Sect. 2.3.2.2), given that

these are the assumptions where stewardship theory and principal-agent theory

most starkly differ (see Sect. 2.3.3.1).83 Specifically, stewardship theorists have

defined a number of psychological characteristics, perceptions of the organisation’s

situational characteristics, and expectations (see below) that can influence whether

the manager’s natural predisposition to performance maximisation is as assumed in

principal-agent theory or as in stewardship theory (e.g. Davis et al. 1997b,

pp. 27–43; Lane et al. 1999, p. 1079; Vargas Sanchez 2001, 2004, 2005).84

As these variables shift the manager’s natural predisposition to performance

maximisation towards those of an agent (i.e. assumption 1: self-serving utility

maximisation and assumption 4: less performance-maximising interests than the

principal), the impact of discretion on performance becomes more negative, as

explained in Sect. 2.3.2.2. By contrast, as these variables shift towards stewardship

(i.e. assumption 1: pro-organisational utility maximisation and assumption 4:

equally or more performance-maximising interests than the principal), the impact

of discretion on performance becomes more positive. Consequently, by affecting

the moderator of the manager’s natural predisposition to performance

maximisation, the specified psychological characteristics, perceptions of theorganisation’s situational characteristics, and expectations can moderate the

impact of managerial discretion on performance.85 These moderator variables

83 Initial research on stewardship theory contrasted principal-agent theory and stewardship theory

(e.g. Donaldson and Davis 1989, 1991, 1994; Fox and Hamilton 1994) and tended to assume that

stewardship theory is correct and principal-agent theory is incorrect (see Davis et al. 1997b, p. 21).

The seminal research by Davis et al. (1997b) from which moderating effects are derived herein no

longer assumes that only stewardship theory is correct but rather attempts to reconcile the two

theories’ differences by delineating the conditions (i.e. the moderator variables) under which each

of the theories is necessary.84 As explained above in the discussion on antecedents, stewardship theorists refer to these

variables as ‘antecedents in the prediction of stewardship versus agency relationships’ (Davis

et al. 1997b, p. 37).85 If the choice between stewardship and agency behaviour (which is argued to depend on these

variables) is viewed as a binary choice, then these variables represent moderator variables that

switch between a positive (stewardship) and a negative (agency) impact of discretion on perfor-

mance. If instead the choice is viewed as lying on a continuum between stewardship and agency

behaviour (e.g. Albanese et al. 1997, p. 610), then these variables can have continuous moderating

effects. The more the variables match the assumptions of stewardship rather than agency theory,

the more positive the impact of discretion on performance is predicted to be. As these moderators

work via altering the manager’s natural predisposition to performance maximisation, they may

moderate the impact of other moderators (e.g. monitoring control mechanisms)—which was

discussed in Sect. 2.3.2 in terms of three-way interaction (e.g. Henseler and Fassott 2010,

p. 722) and enter into the study’s new discretion model in Chap. 4.


(i.e. influences on managers)—which are indicated in Fig. 2.5 above as influences

on whether discretion’s impact on performance is an upwards-sloping line

(i.e. positive impact for a steward) or a downwards-sloping line (i.e. negative

impact for an agent)—have specifically been defined as follows:

• Psychological characteristics. Davis et al. (1997b, pp. 27–38) argue that the

manager’s psychological characteristics—his/her motivation, identification, and

use of power—influence whether a manager is likely to be a self-serving utility

maximiser (i.e. agent) or a pro-organisational, collective self-actualiser who

achieves utility through organisational achievement (i.e. steward). As to motiva-tion, managers motivated by extrinsic lower-order economic needs are argued to

be more likely to behave as agents, whereas managers aspiring to intrinsic

higher-order needs of self-actualisation are more likely to act as stewards

(e.g. Argyris 1973a, b; Khanchel 2009, p. 98; Maslow 1970; McGregor 1960;

Simon 1957a, b). Identification, i.e. the extent to which the organisation is an

extension of the manager’s psychological structure, is likewise considered.

Managers with low value commitment are deemed more susceptible to agency

behaviour, while those with high value commitment favour stewardship

behaviour (e.g. Brown 1969; Caza 2007; Katz and Kahn 1978; Khanchel 2009,

p. 98; Turner 1981). Finally, as to the manager’s use of power, managers

emphasising institutional power (i.e. legitimate, coercive, and reward power)

are more prone to agency behaviour and managers stressing personal power (i.e.

expert and referent power) are seen as more likely to exhibit stewardship

behaviour (e.g. Davis et al. 1997b; French and Raven 1959; Gibson et al.

1991; Khanchel 2009, p. 98).

• Perceptions of the organisation’s situational characteristics. Both an

organisation’s management philosophy and cultural differences are postulated

to act as further influences on whether a principal-manager relationship exhibits

stewardship versus agency behaviour (Davis et al. 1997b, pp. 32–38). A control-

oriented management philosophy tends to support principal-agent theory,

whereas an involvement-oriented management philosophy supports stewardship

theory (e.g. Lawler 1986, 1992; Walton 1980, 1985). A control-oriented man-

agement philosophy entails a risk orientation of control mechanisms rather than

trust, a short-term time frame rather than a long-term time frame, and cost

control rather than performance enhancement objectives (Davis et al. 1997b,

pp. 32–37; Van Slyke 2007, p. 164). As to cultural differences, individualismand high power distance make it more likely that the principal-agent theory’s

assumptions apply whereas collectivism and low power distance make the

stewardship theory’s assumptions more probable (e.g. Hofstede 1980, 1991;

Triandis 1990, 1995; Triandis et al. 1993; Vargas Sanchez 2004). These cultural

differences are further discussed in Sect. 2.4.2, as they contribute to the formu-

lation of Hypotheses 2 and 3.

• Expectations (e.g. risk and trust). Instead of postulating that these psychologicaland situational characteristics decide in a deterministic way whether a relation-

ship is an agency relationship or a stewardship relationship, Davis et al. (1997b,


pp. 38–43) argue that these characteristics act as inputs into the principal’s and

manager’s decisions for choosing agency versus stewardship behaviour. This

choice resembles a prisoner’s dilemma in game theory (Gibbons 1992;

Vargas Sanchez 2004, pp. 4–5, 2005, pp. 24–25; von Neumann and Morgenstern

1944) and therefore each party’s expectations of the other party play an impor-

tant role in determining the outcome (i.e. an agency relationship or a stewardship

relationship). In particular, it is explained in Box 2.4 that the principal’s and

manager’s choice of the type of principal-manager relationship may depend on

the parties’ expectations, including their willingness to trust each other and theirappetite for risk (e.g. Davis et al. 1997b; Mills and Keast 2009; Van Slyke 2007).

Box 2.4: Game for Choosing Between Agency and Stewardship Relationships

This Box explains the game proposed by Davis et al. (1997b, pp. 38–43) for

representing the principal’s and the manager’s choice of the nature of their

relationship. It highlights the role of expectations in determining whether a

relationship conforms to stewardship theory (whereby discretion increases

performance) or principal-agent theory (whereby discretion decreases perfor-

mance). In line with game theory (Gibbons 1992, p. 1), this game is presented

below in terms of players choosing actions that yield payoffs which depend

on their combined actions:

1. Players. The two players are the principal and his/her subordinate manager

(i.e. agent or steward). Their own psychological characteristics, their

perceptions of the organisation’s situational characteristics, and their

expectations of the other party (all three of which were described above)

act as inputs into the principal’s and manager’s decisions for choosing

agency versus stewardship behaviour.

2. Actions. The principal chooses whether to empower the manager with

more discretion (stewardship behaviour) or to impose controls to limit the

discretion of the manager (agency behaviour). Likewise, the manager

chooses whether to use the resulting discretion to the benefit of the

organisation (discretion increases performance; stewardship behaviour)

or to use discretion at the expense of the organisation (discretion decreases

performance; agency behaviour).

3. Payoffs. Due to the two alternative behaviours by the two parties, there are

four possible types of relationships: a true principal-steward relationship

(in which the potential joint performance is maximised), a true principal-

agent relationship (in which each party’s expectations are likely to be

achieved and thus the losses to each party minimised), and two mixed-

motive relationships (in which one party chooses stewardship behaviour

and the other chooses agency behaviour). For the two mixed-motive

relationships, the party choosing stewardship behaviour feels betrayed

by the other opportunistically-acting party, making the betrayed party


prefer the true principal-agent relationship to this mixed-motive relation-

ship.86 Hence, if either party expects the other party to choose agency

behaviour, its best response is to also choose agency behaviour so as to

minimise its losses, which leads to a true principal-agent relationship with

suboptimal performance (Davis et al. 1997b, p. 43).

Although as described above, the psychological characteristics and

perceptions of the organisation’s situational characteristics predispose

whether a party is inclined to choose agency behaviour (individualisticorientation) or stewardship behaviour (collectivist orientation), the resultingrelationship also hinges critically on the parties’ expectations of one another

(Davis et al. 1997b, p. 40):

• When both the principal and the manager have individualistic orientations,the resulting relationship inevitably is a true principal-agent relationship,where both parties choose agency behaviour. The reason is that for an

individualistic party, the best choice is agency behaviour regardless of the

behaviour of the other party, since it allows taking advantage of the other

party if the other party chooses stewardship behaviour andminimises losses

if the other party chooses agency behaviour.

• When both the principal and the manager have collectivist orientations, theresulting relationship depends on the parties’ expectations, including their

willingness to trust each other and appetite for risk. As the collectivist

parties subordinate their personal goals to collective goals and evaluate

their joint utility, they would achieve the highest utility by both choosing

stewardship behaviour, whereby the principal grants substantial discretion

and the steward uses this discretion to benefit the organisation. The best

response to stewardship behaviour by one party is thus stewardship

behaviour by the other party. However, if one party chooses agency beha-

viour, the best response for the other party is no longer stewardship

behaviour but also to choose agency behaviour so as to minimise losses

(see above). Therefore, if a collectivist party expects the other party to

choose stewardship behaviour, its best response is also stewardship

behaviour, while if it expects the other party to choose agency behaviour,

its best response is agency behaviour. As it is difficult to know ex ante

whether or not the other party is trustworthy and will engage in stewardship

behaviour, it is prudent to choose agency behaviour so as to limit(continued)

86 For example, suppose the manager acted as a steward but was controlled by the principal as if

he/she were an agent, e.g. by harshly constraining the steward’s discretion through control

mechanisms. The steward might then feel demotivated due to the theoretical mechanism related

to motivation described earlier and potentially engage in antiorganisational behaviour (Argyris

1964; Fleishman and Harris 1962; Herzberg et al. 1959). Likewise, in the other type of mixed-

motive relationship, the ‘principal is likely to feel betrayed and angry and may increase controls,

withdraw from the situation, or attempt to remove the manager’ (Davis et al. 1997b, p. 40).


the potential losses from betrayal in the mixed-motive relationships

(Davis et al. 1997b, p. 22; Williamson 1985). Consequently, the principal’s

and the manager’swillingness to trust one another and their appetite for riskwill influence whether the parties act in line with stewardship behaviour

versus agency behaviour.

In sum, whether a principal-agent or a principal-steward relationship

results depends not only on the parties’ own psychological characteristics

and their perceptions of the organisation’s situational characteristics, but also

on their expectations of the other party, including their willingness to trust

each other and their appetite for risk (e.g. Davis et al. 1997b; Mills and

Keast 2009; Van Slyke 2007).

It should be noted that although this game is formulated as a static rather

than a dynamic game (i.e. consists of a choice between agency and steward-

ship behaviour at a single point in time), stewardship theorists have qualita-

tively added dynamic aspects to the choice between agency and stewardship

relationships. In particular, it has been argued that the parties can invest

in developing trustworthy relations over time (e.g. Van Slyke 2007,

pp. 165–166) and that if a mixed-motive relationship is initially chosen

where one party is betrayed, then the relationship will inevitably progress

towards a principal-agent relationship (e.g. Davis et al. 1997b, p. 41).

On the one hand, this combination of principal-agent theory and stewardship theory

(proposed by stewardship theorists; e.g.Davis et al. 1997b, pp. 27–43;Lane et al. 1999,

p. 1079; Vargas Sanchez 2001, 2004, 2005) takes a further step towards resolving the

discretion puzzle and is therefore integrated into the present study’s new theoretical

discretion model (see Sect. 4.1.2) as well as into its Hypotheses 2 and 3 (see

Sect. 2.4.2).87 As noted earlier in this section on the discretion puzzle, a reason why

existing empirical evidence on the impact ofmanagerial discretion onperformance has

ranged from positive (d > 0) (e.g. Agarwal et al. 2009; Barnabas and Mekoth 2010;

Chang andWong 2003; Gammelgaard et al. 2010; Khanchel 2009) to neutral (d � 0)

(e.g. Caza 2011; Groves et al. 1994; Yougen Li and Zhao 2004; Lopez-Navarro and

Camison-Zornoza 2003; Venaik 1999) and even negative (d < 0) (e.g. He et al. 2009;

Heinecke 2011; Stano 1976;Williamson 1963; Xu et al. 2005) might be due to certain

studies predominantly sampling agents as the unit of analysis where the assumptions

of principal-agent theory apply (d < 0), others focusing on stewards where the

87 Section 1.1 on the research gap mentions this combination of principal-agent theory and

stewardship theory as one of the existing attempts to reconcile the relevant theories. The other

attempts mentioned in Sect. 1.1 relate to combinations of managerial discretion theory and

principal-agent theory (e.g. Caza 2007; 2011; Finkelstein and Peteraf 2007), which have been

discussed in Sect. 2.3.2.


assumptions of stewardship theory apply (d > 0), and yet others mixing agents and

stewards (d � 0). The key differences in the assumptions of principal-agent theory and

stewardship theory were shown to pertain to the manager’s natural predisposition to

performance maximisation (i.e. assumption 1: the manager’s utility function and

assumption 4: the manager’s interest in maximising performance relative to the

principal). It was further shown that the extent to which the theories’ assumptions on

human behaviour apply might depend on the principal’s andmanager’s psychological

characteristics, perceptions of the organisation’s situational characteristics, and

expectations. Consequently, taking a more granular approach (as in the present

study) that differentiates the unit of analysis according to such moderator variables

might reveal that particular profiles of psychological characteristics, situational

characteristics, and expectations can lead to a positive (d > 0) and other profiles to a

negative (d < 0) impact of discretion on performance. These moderator variables are

built upon inSect. 2.4.2 in order towork towards explaining the contradictory evidence

of the discretion puzzle.

On the other hand, however, while these psychological characteristics, situational

characteristics, and expectations might help resolve the discretion puzzle (and are

therefore integrated into this study’s theoretical discretion model and hypotheses),

they appear to be only part of the puzzle’s resolution for a number of reasons:

• As this reconciliation is based on the individuals’ psychologies, which are difficult

to evaluate both in empirical research and in practice, it is inherently difficult toverify empirically and difficult to apply in practice. Even if it were empirically

proven that certain psychological characteristics, perceptions, and expectations

predicted whether a given manager acted as a steward versus an agent,88 it would

be intricate for a principal in practice to observe these factors and thus evaluate ex

ante whether or not the manager would act as a steward or an agent (Davis et al.

1997b, p. 22; Williamson 1985). Provided the principal (e.g. top management)

cannot sufficiently evaluate the psychology of themanager (e.g. the plant manager),

88 A subset of these variables has been empirically investigated (e.g. Caza 2007; 2011; Mills and

Keast 2009; Van Slyke 2007; Vargas Sanchez 2001, 2004). For agrarian cooperative societies in

Spain (Vargas Sanchez 2001) and Wales (Vargas Sanchez 2004), for example, a subset of the

psychological characteristics and situational characteristics are empirically confirmed as deter-

mining agency versus stewardship relationships of chairmen and managers. For Spain, individuals

motivated by higher-order needs as well as intrinsic needs were found to be more likely to develop

steward relationships, as were individuals who tended to use personal power rather than institu-tional power (Vargas Sanchez 2001). For Wales, by contrast, stewardship relationships were

predicted by greater identification with the organisation and to a lesser extent by a lower powerdistance (Vargas Sanchez 2004). The other tested factors, such as value commitment,

involvement-orientation, and collectivist cultures received no empirical support in either study.

In addition, Caza (2007) tests for moderating effects of managerial commitment (i.e. identifica-tion; see the psychological characteristics above) on the impact of discretion on performance, but

finds no significant evidence. By contrast, Van Slyke (2007) finds a positive association between

the risk that either of the parties perceives and the extent to which agency behaviour is adopted (seethe expectations above).


this reconciliation (while potentially true theoretically) might thus not be able to

recommend whether to increase or decrease managerial discretion in practice.

• Even if these psychological factors could determine stewardship versus agency

behaviour of amanager, they could at best explain whether or not a givenmanager

will choose to work diligently towards the organisation’s performance and not

whether the manager’s actions will strengthen or weaken performance, which can

also depend on such moderator variables as the manager’s ability, knowledge or

information (see Box 2.2 in Sect. 2.3.2.2; e.g. Caza 2007, p. 13, Caza 2011, p. 9;

Chang and Wong 2003, p. 24; Davis et al. 1997b, pp. 23–24; Xiaoyang Li 2007).

• Finally, this reconciliation does not integrate managerial discretion theory (see

Sect. 2.3.1). In particular, the potential multidimensionality of discretion is

overlooked (e.g. Carpenter and Golden 1997, p. 195; Caza 2007, pp. 26–82;

Chen 2006; Finkelstein and Peteraf 2007, p. 245; Hambrick and Abrahamson

1995, p. 1439; Hambrick and Finkelstein 1987, pp. 371–402; Hambrick et al.

1993, p. 409; see Sect. 2.1.2). By integrating all three relevant theories into a

new empirically-validated model in Chap. 4, this study demonstrates that the

above reconciliation postulating that a manager is either a steward or an agent

cannot fully explain the impact of discretion on performance (see Sect. 7.4.2;

also see Sect. 2.4.3 on Hypothesis 4).

The reviews of principal-agent theory and stewardship theory (see Sects. 2.3.2

and 2.3.3, respectively) have provided a theoretical explanation for why different

units of analysis might exhibit different impacts of discretion on performance. This

explanation posits that different managers analysed (i.e. different units of analysis)

might be subject to different influences (i.e. moderator variables) and that these

influences affect how these managers use their discretion and thereby whether

discretion tends to increase or decrease performance. The influences derived from

principal-agent theory include the manager’s natural predisposition to performance

maximisation, compensation control mechanisms, the nature of managerial activities,

monitoring control mechanisms, and the manager’s ability, knowledge, and informa-

tion (see Table 2.9 in Sect. 2.3.2.2). Likewise, stewardship theorists have argued that

the principal’s and manager’s psychological characteristics, perceptions of the

organisation’s situational characteristics, and expectations may act as potential

influences (i.e. moderators) on whether managers use their discretion so as to

increase or decrease performance (see Fig. 2.5 above). These influences are translated

into two moderators in Sect. 2.4.2, namely firm type (Hypothesis 2) and firm size

(Hypothesis 3). By testing whether the impact of discretion on performance (@P @D= )

can differ by these moderator variables (M), the present study establishes whether

managers under different influences (i.e. different firm types and firm sizes) tend to

use their discretion in different ways (i.e. more like stewards or more like agents) and

thus whether the moderators help predict a positive versus negative impact of

discretion on performance—i.e. work towards resolving the discretion puzzle.

In sum, following the literature reviews of managerial discretion theory (see

Sect. 2.3.1) and principal-agent theory (see Sect. 2.3.2), this section has reviewed


the literature on stewardship theory briefly in general terms (e.g. historical context,

unit of analysis, assumptions, and predictions) as well as thoroughly regarding the

impact of discretion on performance. The literature review summary Table 2.11


Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality







Impact on

performance:

• Direct

• Moderating

Contradictory

evidence

• Positive,

neutral,

negative

• Moderators

potentially

important

but rarely

modelled

(esp. firm

type and

firm size)

Theoretical

• Not specified

• Not specified.

Antecedents:

environmental,

organisational,

and managerial

Mechanisms

specified

• Negativea

• E.g. agent’s

interests relative

to principal’s

and asymmetry

of information

Mechanisms

specified

• Positive

• Psychological

and situational

characteristics

as well as

expectations


management,

rarely China

Extended from

top to middle

management

Extended from

top to middle

management

Extended from

top to middle

management

Section 2.4

Hypothesis 1

Middle

management

in China

Hypothesis 2

Hypothesis 3

Hypothesis 4

context specified









reveals that all three theories apply to middle management and thus to this study’s

unit of analysis. Moreover, similarly to principal-agent theory and in contrast to

managerial discretion theory, stewardship theory is found not to focus on specifying

the theoretical context of managerial discretion in terms of the construct and

antecedents of discretion, but rather to focus on the consequences of discretion on

performance: Stewardship theory makes assumptions regarding human behaviour

that differ starkly from those in principal-agent theory and imply theoretical

mechanisms which predict a positive impact of discretion on performance (as

noted in Table 2.11). In addition to these direct effects, stewardship theorists

have postulated a number of moderating effects based on the principal’s and

manager’s psychological characteristics, perceptions of the organisation’s situa-

tional characteristics, and expectations. These direct and moderating effects as well

as other findings from the literature review are synthesised into various implications

for the present study in the next section.

2.4 Implications of Literature Review

The preceding sections of this chapter have conducted a thorough, in-depth review

of both the empirical and the theoretical literature on the impact of managerial

discretion on performance. This section synthesises the review into implications for

the present study that fulfil the four purposes of the literature review described at the

outset of this chapter:

1. Deriving the research gap and research objective (see Sects. 1.1 and 1.2).

2. Formulating the postulate and hypotheses (see Sects. 2.4.2 and 2.4.3).

3. Selecting the unit of analysis (see Chap. 3).4. Developing the study’s new discretion model (see Chap. 4).

The second of these purposes, i.e. formulating the postulate and hypotheses, is

the main focus of this section. The other three purposes are addressed only briefly,

as they are further scrutinised in other chapters (see Chap. 1 for the research gap and

research objective, Chap. 3 for the unit of analysis, and Chap. 4 for the study’s new

discretion model).

The way in which the present section derives implications for these four

purposes is by analysing Table 2.12, which reinserts the literature review summary

table that has been incrementally compiled throughout the previous sections. As the

table depicts the current state of the empirical and theoretical literature (see the

columns), it can be synthesised across the columns into the research gap and

research objective (first purpose; see Sect. 2.4.1), the postulate and hypotheses

(second purpose; see Sects. 2.4.2 and 2.4.3), and the unit of analysis (third purpose;

see Sect. 2.4.4). Moreover, the completed review allows Chap. 4 to develop a new


discretion model (fourth purpose; see Sect. 2.4.5) in a way that integrates the

empirical and the theoretical literature and permits the study to test its four

hypotheses for its unit of analysis so as to fulfil its research objective.


Empirical

evidence

Managerial

discretion

theory

Principal-

agent theory

Stewardship

theory

Implications

for this study

Discretion:

• Definition

• Dimensionality







Impact on

performance:

• Direct

• Moderating

Contradictory

evidence

• Positive,

neutral,

negative

• Moderators

potentially

important

but rarely

modelled

(esp. firm

type and

firm size)

Theoretical

context specified

• Not specified

• Not specified.

Antecedents:

environmental,

organisational,

and managerial

Mechanisms

specified

• Negativea

• E.g. agent’s

interests

relative to

principal’s

and asymmetry

of information

Mechanisms

specified

• Positive

• Psychological

and situational

characteristics

as well as

expectations


management,

rarely China

Extended from

top to middle

management

Extended from

top to middle

management

Extended from

top to middle

management

Section 2.4

Hypothesis 1

Middle

management

in China

Hypothesis 2

Hypothesis 3

Hypothesis 4








2.4 Implications of Literature Review 121

2.4.1 Implications for Research Gap and Research Objective

This section demonstrates how the research gap and research objective of the presentstudy are derived from the literature review summary in Table 2.12 above.89 When

comparing the results from the empirical literature and theoretical literature in

Table 2.12 with respect to the row ‘Impact on performance: Direct’, the study’s

research gap, which is termed the discretion puzzle (see Sect. 1.1), becomes apparent:90

• The empirical literature (see the column ‘Empirical Evidence’ in Table 2.12)

reviewed in Sect. 2.2.2 has comprised divergent results on the impact of mana-

gerial discretion on performance, ranging from positive (e.g. Agarwal et al.

2009; Barnabas and Mekoth 2010; Chang and Wong 2003; Gammelgaard

et al. 2010; Khanchel 2009) to neutral (e.g. Caza 2011; Groves et al. 1994;

Yougen Li and Zhao 2004; Lopez-Navarro and Camison-Zornoza 2003; Venaik

1999) and even negative (e.g. He et al. 2009; Heinecke 2011; Stano 1976;

Williamson 1963; Xu et al. 2005). This contradictory evidence as to whether

discretion tends to increase, not alter or decrease performance spans different

levels of management (e.g. top management and middle management) as well as

different geographies (e.g. United States of America, European countries, and

China; see Sect. 2.2.2). Even for a given level of management in a given country,

such as top management in China, empirical studies have found positive, neutral,

and negative impacts of managerial discretion on performance (e.g. Chang and

Wong 2003; 2004; Yougen Li and Zhao 2004; Xu et al. 2005; Zhang 1997).

• The theoretical literature (see the columns ‘Managerial Discretion Theory’,

‘Principal-Agent Theory’, and ‘Stewardship Theory’ in Table 2.12) reviewed

in Sect. 2.3 provides no complete answer as to whether discretion tends to

increase, not alter or decrease performance. While managerial discretion theory

(see Sect. 2.3.1)91 specifies the theoretical context of managerial discretion,

the nature of the impact of discretion on performance (e.g. positive versus

negative) remains unspecified. Furthermore, although principal-agent theory

89A thorough discussion of the research gap and research objective can be found in Sects. 1.1 and 1.2.90 As mentioned above, the present study’s discretion puzzle relates to the managerial discretion of

managers in organisations and is therefore distinct from the puzzle of discretion (Pratt and Sossin

2009) that concerns judicial discretion in law.91Managerial discretion theory (e.g. Abrahamson and Hambrick 1997; Berman et al. 2005;

Carpenter and Golden 1997; Caza 2007; 2011; Huiyuan Chen 2006; Crossland 2007;

Crossland and Hambrick 2007; Datta et al. 2003; Finkelstein and Boyd 1998; Finkelstein and

Hambrick 1990; Finkelstein and Peteraf 2007; Haleblian and Finkelstein 1993; Hambrick

and Abrahamson 1995; Hambrick and Finkelstein 1987; Hambrick et al. 1993; Hutzschenreuter

and Kleindienst 2007; Keegan 2006; Keegan and Kabanoff 2008; Key 2002; Yougen Li and Zhao

2004; Magnan and St-Onge 1997; Quigley and Hambrick 2009; Rajagopalan and Finkelstein

1992; Thomas and Peyrefitte 1996; Zhang and Li 2008b; Zhang et al. 2006a, b).


(see Sect. 2.3.2)92 and stewardship theory (see Sect. 2.3.3)93 specify mechanisms

for the impact of discretion on performance, principal-agent theory predicts the

impact to be negative whereas stewardship theory predicts it to be positive. None

of these existing theories can therefore fully account for why certain empirical

studies find a positive, others a neutral, and yet others a negative impact of

discretion on performance—neither when treating the three relevant theories

individually, nor when treating them collectively in existing theory

combinations (see Sect. 2.3; e.g. Caza 2007; 2011; Davis et al. 1997b,

pp. 27–43; Finkelstein and Peteraf 2007, pp. 237–243; Lane et al. 1999, p.

1079; Vargas Sanchez 2001, 2004, 2005).

From this synthesis, it can now clearly be seen that the empirical and theoretical

literature in Table 2.12 together generate the present study’s research gap termed the

discretion puzzle (see Sect. 1.1): The ostensible paradox that empirical evidence

coexists for positive, neutral, and negative impacts of discretion on performance,

which the existing theories do not unequivocally explain, neither individually nor

collectively. Hence, in order to narrow this research gap, the present study’s researchobjective (see Sect. 1.2) is set to work towards resolving the discretion puzzle createdby the contradictory empirical evidence on the impact of managerial discretion on

performance that cannot be fully explained by the existing theories (i.e. managerial

discretion theory, principal-agent theory, and stewardship theory).94

2.4.2 Implications for Postulate and Hypotheses:Hypotheses 1 to 3

The study’s review of the empirical and the theoretical literature summarised in

Table 2.12 above allows the present study to formulate its postulate and hypothesesfor working towards resolving the discretion puzzle. Hypotheses 1, 2, and 3 are

92 Principal-agent theory (e.g. Agrawal and Knoeber 1996; Baysinger and Butler 1985; Berger

et al. 1997; Brush et al. 2000; Chang and Wong 2003; Childs and Mauer 2008; Denis et al. 1997;

Eisenhardt 1989; Fama 1980; Fama and Jensen 1983a, b; He et al. 2009; Jensen 1986; Jensen and

Meckling 1976; Jensen and Murphy 1990; Jensen and Ruback 1983; Laffont and Martimort 2002;

Lang et al. 1995; Levinthal 1988; Ongore 2011; Shleifer and Vishny 1997; Spremann 1987;

Thepot 2007; Thomsen and Pedersen 2000; Walters 1995; Wang et al. 2008; Weidenbaum and

Jensen 1993; Werner and Tosi 1995, p. 1673; Xu et al. 2005; Zou 1989).93 Stewardship theory (e.g. Albanese et al. 1997; Arthurs and Busenitz 2003; Corbetta and Salvato

2004; Davis et al. 1997a, b; Dicke and Ott 2002; Donaldson 1990; Donaldson and Davis 1989,

1991, 1993, 1994; Eddleston and Kellermanns 2007; Fox and Hamilton 1994; Lane et al. 1999; Liu

and Cai 2004; Miller and Le Breton-Miller 2006; Mills and Keast 2009; Muth and Donaldson

1998; Salvato 2002; Tian and Lau 2001; Tosi et al. 2003; Van Slyke 2007; Vargas Sanchez 2001,

2004, 2005; Zahra 2003).94More specifically, the research objective is to establish that the failure of the extant literature to

account for granularity in the way that managers use discretion is a potential cause of the discretion

puzzle—and that theories and empirical studies must therefore differentiate discretion’s impact by

this granularity (i.e. by dimensions of discretion and influences on managers) to resolve the

discretion puzzle (see Sect. 1.2).


developed by combining the empirical and theoretical literature in three steps.

These steps are briefly outlined below and described in greater detail thereafter:

1. The review of the empirical literature (see Sect. 2.2) has diagnosed that potentialcauses for the differences in the empirical results of existing studies (i.e.

positive, neutral, and negative impacts) are the differences in the research

designs of these existing studies, such as different measures of discretion and

different units of analysis. As explained below, this empirically motivates the

postulate that the impact of discretion on performance must be investigated with

greater granularity, i.e. differentiating discretion in a more fine-grained manner

and differentiating the unit of analysis by e.g. firm type and firm size.

2. The review of managerial discretion theory (see Sects. 2.1 and 2.3.1) has

developed a theoretical explanation for why different measures of discretion

might produce contradictory impacts on performance in empirical studies.

Discretion might be multidimensional, with managers using discretion in differ-

ent areas of their work (i.e. dimensions) differently. Different dimensions and

thus measures of discretion could then have different impacts on performance.

As described below, this leads to Hypothesis 1 (see Box 2.5).

3. The reviews of principal-agent theory and stewardship theory (see

Sects. 2.3.2 and 2.3.3) have led to a theoretical explanation of why different

units of analysis might exhibit different impacts of discretion on perfor-

mance. The managers analysed might be subject to different influences

(e.g. firm type and firm size; see Box 2.6) that affect how the managers use

their discretion and thereby moderate discretion’s impact on performance. As

discussed below, this motivates testing Hypothesis 2 and Hypothesis 3 (see

Box 2.5).

Box 2.5: Postulate and Hypothesis 1, Hypothesis 2, and Hypothesis 3

Postulate of the Present StudyManagers may use managerial discretion (i.e. the latitude of managerial

action) differently depending on the area of their work in which discretion

is granted (i.e. the dimension of discretion) and the influences on managers in

place (e.g. firm type and firm size).95 The impact of managerial discretion on

performance can therefore differ by the dimensions of discretion, firm type,

and firm size. Hence, theories and empirical studies must differentiate the

impact of discretion by greater granularity (e.g. by dimensions, firm type, and

firm size) in order to work towards resolving the discretion puzzle.

95 For example, a given manager might use additional discretion for making capital investments in

a way that improves performance but use additional discretion for hiring workers in a way that

reduces performance. According to this postulate, discretion could have positive, neutral, and also

negative impacts on performance depending on the dimensions of discretion (e.g. capital invest-

ment discretion versus hiring discretion), firm type (e.g. Chinese firms versus multinationals), and

firm size (e.g. 150 versus 5,000 employees).





Hypothesis 2 (Firm Type)

H02: The impact ofmanagerial discretion on performance is equal between different


H12: The impact ofmanagerial discretion on performance differs between different


Hypothesis 3 (Firm Size)

H03: The impact ofmanagerial discretion on performance is equal between different

firm sizes.


firm sizes.

The reviews of the empirical literature, managerial discretion theory, principal-

agent theory, and stewardship theory therefore collectively define Hypotheses 1 to

3, which culminate in the present study’s postulate described in Box 2.5: Managers

may use managerial discretion differently depending on the area of their work in

which discretion is granted (i.e. Hypothesis 1: dimension of discretion) and the

influences on managers in place (i.e. Hypothesis 2: firm type; Hypothesis 3: firm

size; see footnote 95 of this chapter). The impact of discretion on performance

could then differ by the dimensions of discretion, firm type, and firm size—and

theories and empirical studies would thus need to differentiate the impact of

discretion by this greater granularity (i.e. dimensions, firm type, and firm size) in

order to work towards resolving the discretion puzzle. The remainder of this section

explains the study’s postulate and Hypotheses 1, 2, and 3 in greater detail by

following the three steps outlined above.

The first step revisits the review of the current state of the empirical literature inSect. 2.2, which has motivated the study’s postulate in Box 2.5 to differentiate the

impact of discretion with greater granularity:

• While existing studies have varied strongly in their empirical results on the

estimated impact of managerial discretion on performance (ranging from posi-

tive to neutral and even negative; see the discretion puzzle above), they have also

varied strongly in their research designs (e.g. measure of discretion and unit of

analysis; e.g. Adams et al. 2005; Huiyuan Chen 2006; Cheng et al. 2006;


Heinecke 2011; Thomas and Peyrefitte 1996; Walters 1995; Wang et al. 2008;

Xu et al. 2005; Yan et al. 2010; Zhang and Li 2008b). The differences in research

designs might therefore constitute potential causes for the differences in the

estimated results for the impact of managerial discretion on performance.96

• Yet as research designs have simultaneously differed in many ways between

existing studies (e.g. Agarwal et al. 2009; Barnabas andMekoth 2010; Caza 2011;

Crossland and Hambrick 2007; Khanchel 2009; Yougen Li and Zhao 2004;

Lopez-Navarro and Camison-Zornoza 2003; Ongore 2011; Tang 2008;

Williamson 1963), it is difficult to infer from the existing studies whether any

particular difference in the research design (e.g. the measure of discretion) might

explain the divergent empirical results. The solution proposed by the present

study is to test whether an individual potential cause (i.e. a difference in the

research design identified from comparisons between studies) produces differentempirical results within a single empirical study, since this holds all other

potential causes (i.e. other aspects of the research design) constant.

• Consequently, this study develops a new empirical discretionmodel (see Chap. 4)

that exhibits greater granularity than the models in many previous studies (e.g.

Bowen et al. 2008; Chang and Wong 2003; Gammelgaard et al. 2010;

Groves et al. 1994; He et al. 2009; Xiaoyang Li 2007; Stano 1976; Venaik 1999;

Werner and Tosi 1995; Zhang 1997), i.e. a model that differentiates the measure

of discretion by dimensions of discretion and the unit of analysis by firm type and

firm size. This makes it possible to test the hypotheses in Box 2.5 that different

dimensions of discretion (and thus measures of discretion) as well as different

firm types and firm sizes (and thus units of analysis) can lead to different impacts

of discretion on performance (i.e. different empirical results).97 In other words,

it empiricallymotivates the postulate that the impact of discretion on performance

must be investigated with greater granularity, i.e. differentiating discretion by

dimensions (Hypothesis 1) and differentiating the unit of analysis by firm type and

firm size (Hypotheses 2 and 3).

96 For example, if two different studies adopt two different measures of discretion, each of which

taps into a distinct dimension of discretion, then one study might find a positive and the other a

negative impact of discretion on performance, provided the distinct discretion dimensions have

different performance impacts.97 It should be noted that as explained in Box 1.1 in Sect. 1.2 on the delimitations of the research

objective, the present study’s approach is designed as a proof-by-counter-example and is subject to

the caveat of observational cross-sectional studies in terms of demonstrating causality (e.g. Caza

2007, p. 46; Finkelstein and Hambrick 1990, p. 500; Granger 1969; Sanchez 2008, p. 5; Simon

1954, pp. 477–478; Wagner 2002, pp. 287–292; see Sect. 5.4.1). In simple terms, if in the instance

of the present study it is found that discretion is multidimensional, this would constitute a proof-

by-counter-example that discretion is not unidimensional in all cases. While this does not prove

that unidimensional measures of discretion are flawed in the existing studies and responsible for

the discretion puzzle, it would offer evidence that this is a possibility and warrants further

investigation.


In short, by transferring the differences in research designs between existing

empirical studies into a more granular model within a new empirical study, this

study tests whether differences in research designs (i.e. dimensions of discretion,

firm type, and firm size) might be responsible for differences in estimated impacts

of discretion on performance. It thereby tests whether greater granularity can help

explain the positive, neutral, and negative impacts of discretion that give rise to the

discretion puzzle.

The second step derives Hypothesis 1 from the empirical and theoretical litera-

ture (summarised in Table 2.12 above; see the row ‘Discretion’), which leads to the

postulate that managers may use managerial discretion (i.e. the latitude of manage-

rial action) differently depending on the area of their work in which discretion is

granted (i.e. the dimension of discretion; see Box 2.5):

• The preceding analysis of the empirical literature has suggested that differencesin the measures of discretion between existing empirical studies might poten-

tially help explain the differences in the empirical results between these studies,

which have ranged from positive to neutral and even negative estimated impacts

of discretion on performance (see Sect. 2.2).

• Moreover, based on the theoretical analysis in Sect. 2.1.2 (grounded mostly in

managerial discretion theory; see Sect. 2.3.1), there is reason to believe that

managerial discretion consists of several types, i.e. is multidimensional (e.g.

Barnabas and Mekoth 2010; Carpenter and Golden 1997, p. 195; Caza 2007,

pp. 26–82; Chen 2006; Finkelstein and Peteraf 2007, p. 245; Groves et al. 1994,

p. 190; Hambrick and Abrahamson 1995, p. 1439; Hambrick and Finkelstein

1987, pp. 371–402; Hambrick et al. 1993, p. 409; March and Simon 1958;

Perrone et al. 2003, pp. 422–423).

• Combining these theoretical and empirical points of view, it is conceivable that

discretion is multidimensional in that discretion granted in a certain area of the

manager’s work (e.g. making capital investments) improves performance while

discretion in another domain (e.g. hiring workers) reduces performance—and

that a reason why empirical studies with different measures of discretion have

produced different results might be that their measures have tapped into distinct

dimensions of discretion.

This proposition can be empirically verified by testing Hypothesis 1 (see Box

2.5), namely testing whether discretion is unidimensional (i.e. null hypothesis H01)

versus multidimensional (i.e. alternative hypothesis H11 ) in its impact on perfor-

mance (i.e. testing whether the impact of discretion on performance differs between

dimensions of discretion).

It should be noted that the dimensionality of discretion (i.e. unidimensionality

versus multidimensionality) is still unresolved in the theoretical and empirical

literature (see Sect. 2.1.2), which makes testing Hypothesis 1 a novel contribution

towards resolving the discretion puzzle. Although managerial discretion theoryhints that discretion may be multidimensional (see Sect. 2.1.2), both principal-agent theory (see Sect. 2.3.2) and stewardship theory (see Sect. 2.3.3) tend to treat

discretion as unidimensional (e.g. Chang and Wong 2003; Dicke and Ott 2002,


p. 468; Fox and Hamilton 1994, p. 74; He et al. 2009; Spremann 1987, p. 18;

Vargas Sanchez 2005, p. 19; Xu et al. 2005). Furthermore, many existing empiricalstudies have tended to make the simplifying assumption in their research designs

that discretion is unidimensional and have thus often combined a manager’s

discretion in different areas of his/her work into a single unidimensional discretion

construct (e.g. Barnabas and Mekoth 2010; Bloom et al. 2008; Caza 2007; 2011;

Chang and Wong 2003; Cheng et al. 2006; Gammelgaard et al. 2010; Marin and

Verdier 2006).98 There is thus a discontinuity of postulated multidimensionality (in

managerial discretion theory) and implicitly assumed unidimensionality (in empir-

ical studies, principal-agent theory, and stewardship theory; see Sect. 2.1.2).

Whether or not the assumption of unidimensionality made in many empirical

studies, principal-agent theory, and stewardship theory is universally tenable is

examined by testing the null hypothesis of unidimensionality (H01 ) against the

alternative hypothesis of multidimensionality (H11)—i.e. testing whether the impact

of discretion on performance differs between dimensions of discretion:

• If H01 cannot be rejected, the prevalent simplifying assumption of a uni-

dimensional construct of discretion in existing studies might be justifiable.

• Yet if H01 (i.e. unidimensionality) can be rejected in favour of H1

1 (i.e. multi-

dimensionality), then this simplifying assumption of unidimensionality is not

universally tenable (since it is then not tenable at least in the instance of this

particular study). Rejecting H01 in favour of H1

1 would provide significant

evidence that the performance impact of discretion can depend on the area of

work in which a manager is granted discretion (as postulated in the present study

in Box 2.5). This would help resolve the discretion puzzle by showing that

existing studies might have estimated positive, neutral, and even negative

impacts of discretion on performance because their measures of discretion

have tended not to properly distinguish between the dimensions of discretion,

i.e. the areas of the manager’s work in which discretion is granted.99 More

granular research designs that allowed for the potentially multidimensional

98 Among the many empirical studies resorting to proxy measures for gauging managerial discre-

tion, discretion has prevalently been modelled as unidimensional as well. Empirical studies have

modelled unidimensional discretion constructs by measuring one or several proxies related to e.g.

ratings of managerial power, internal representation on the board of directors, managerial stock

ownership, and financial ratios (e.g. Huiyuan Chen 2006; Khanchel 2009; Yougen Li and Zhao

2004; Zhang and Li 2008b; Zhang et al. 2006a, b) as well as multiple antecedents mainly drawn

from the task environment (e.g. Agarwal et al. 2009; Berman et al. 2005; Cameron 2000;

Finkelstein and Boyd 1998; Finkelstein and Hambrick 1990; He et al. 2009; Magnan and

St-Onge 1997; Rajagopalan and Finkelstein 1992; Williamson 1963). In addition, industry-level

discretion has been frequently proxied in existing studies (e.g. Abrahamson and Hambrick 1997;

Datta et al. 2003; Finkelstein and Hambrick 1990; Hambrick and Abrahamson 1995;

Hambrick et al. 1993; Keegan 2006; Keegan and Kabanoff 2008; Thomas and Peyrefitte 1996).99 As is explained in Chap. 7, erroneously treating discretion as unidimensional when it is in fact

multidimensional might produce misleading estimates of the impact of discretion on performance

that could potentially be a cause of the contradictory empirical evidence that gives rise to the

discretion puzzle.


nature of discretion would then be required in order to produce more meaningful

results in future research.

Put briefly, the review of managerial discretion theory yields a theoretical

explanation for why different measures of discretion might produce contradictory

impacts on performance in empirical studies: Discretion may be multidimensional,

with managers using discretion in different areas of their work differently, so

different dimensions (and thus measures) of discretion could have different impacts

on performance. This proposition is empirically verified by testing Hypothesis 1.

The third step develops Hypothesis 2 (firm type) and Hypothesis 3 (firm size)from the empirical and theoretical literature summarised in Table 2.12 above,

leading to the postulate that managers may use managerial discretion differently

depending on the influences on the managers in place (e.g. firm type and firm size;

see Box 2.5):

• The analysis of the empirical literature in Sect. 2.2 has revealed that studies

finding positive, neutral, and negative impacts of discretion on performance have

often differed in their unit of analysis (e.g. in terms of firm type and firm size;

e.g. Adams et al. 2005; Huiyuan Chen 2006; Cheng et al. 2006; Groves et al.

1994; Heinecke 2011; Lopez-Navarro and Camison-Zornoza 2003; Thomas and

Peyrefitte 1996; Wang et al. 2008; Werner and Tosi 1995; Williamson 1963;

Xu et al. 2005; Yan et al. 2010; Zhang and Li 2008b; Zhang and Xie 2008;

Zhang 1997). This empirically motivates the present study to test whether such

differences in the unit of analysis’ firm type and firm size might moderate the

impact of managerial discretion on performance and thereby help explain why

the impact of discretion has been found to be positive, neutral or negative.

• The reviews of principal-agent theory and stewardship theory (see Sects. 2.3.2and 2.3.3) have provided a theoretical explanation for why different units of

analysis might exhibit different impacts of discretion on performance: Different

managers analysed (i.e. different units of analysis) might be subject to different

influences (i.e. moderator variables) that affect how the managers use their

discretion and thereby whether discretion tends to increase or decrease perfor-

mance. The influences derived from principal-agent theory include themanager’s

natural predisposition to performance maximisation, compensation control

mechanisms, the nature of managerial activities, monitoring control mechanisms,

and the manager’s ability, knowledge, and information (see Table 2.9 in

Sect. 2.3.2.2; e.g. Caza 2007; 2011; Chang and Wong 2003; Cheng et al. 2006;

Davis et al. 1997b; Eisenhardt 1989; Fama and Jensen 1983b; Finkelstein and

Peteraf 2007; Jensen and Meckling 1976; Xiaoyang Li 2007; Wang et al. 2008;

Zhang and Li 2008b). Likewise, stewardship theorists have argued that the

principal’s and manager’s psychological characteristics, perceptions of the

organisation’s situational characteristics, and expectations may act as potential

influences (i.e. moderators) on whether managers use their discretion so as to

increase or decrease performance (see Sect. 2.3.3.2; e.g. Argyris 1973a, b;

Brown 1969; Caza 2007; Davis et al. 1997b, pp. 27–43; French and Raven

1959; Gibson et al. 1991; Hofstede 1980, 1991; Katz and Kahn 1978; Khanchel

2009, p. 98; Lane et al. 1999, p. 1079; Lawler 1986, 1992; Maslow 1970;


McGregor 1960; Mills and Keast 2009; Simon 1957a, b; Triandis 1990, 1995;

Triandis et al. 1993; Turner 1981; Van Slyke 2007, p. 164; Vargas Sanchez 2001,

2004, 2005; Walton 1980, 1985).100

• As these influences on managers (i.e. moderators) may take different values in

different firm types and firm sizes, the present study tests whether differences in

firm type (Hypothesis 2) and firm size (Hypothesis 3) influence the way in which

managers use discretion and thereby moderate the impact of managerial discre-

tion on performance. The reasoning why these influences derived from

principal-agent theory and stewardship theory are expected to differ by the

unit of analysis’ firm type and firm size is discussed in Box 2.6.

In essence, the review of the empirical and theoretical literature leads to the

postulate that managers might be subject to different influences (e.g. firm type and

firm size; see Box 2.5) which affect whether the managers use their discretion to the

benefit of or at the expense of performance. This postulate is empirically verified by

testing Hypothesis 2 and Hypothesis 3 on whether firm type and firm size can affect

the way that managers utilise their discretion to impact on performance.101

If the null hypotheses (H02 and H0

3 in Box 2.5) can be rejected in favour of the

alternative hypotheses (H12 and H

13 in Box 2.5), there would be significant evidence

that the performance impact of discretion can depend on the influences on managers

in terms of firm type and firm size. This would contribute to the resolution of the

discretion puzzle by indicating that existing studies might have estimated positive,

neutral, and even negative impacts of discretion on performance because their units

100 Section 2.3 has derived these influences (i.e. moderators) from the theories’ assumptions, as

these determine the theories’ predicted impact of discretion on performance. For example, the

principal’s and manager’s psychological characteristics, perceptions of the organisation’s situa-

tional characteristics, and expectations are all postulated to influence whether the manager’s

natural predisposition to performance maximisation is as assumed in principal-agent theory (i.e.

discretion is abused to pursue self-serving interests at the expense of performance) or as assumed

in stewardship theory (i.e. discretion is used to diligently improve performance). Empirical studies

focusing on managers as units of analysis where the assumptions of stewardship theory apply (i.e.

stewards) might then find a positive impact of discretion on performance, whereas studies focusing

on agents might find a negative impact, and studies mixing stewards and agents in their samples

might find an insignificant (neutral) impact. It follows that taking a more granular approach that

differentiates the unit of analysis according to moderators related to the theories’ assumptions (i.e.

firm type and firm size in this study) might reveal that certain units of analysis lead to a positive

and others to a negative impact of discretion. This greater granularity could then potentially help

explain the contradictory evidence of the discretion puzzle.101 By testing whether the impact of discretion on performance can differ by these influences (i.e.

firm type and firm size), the present study establishes whether managers in different firm types and

firm sizes tend to use their discretion in different ways (i.e. more like stewards or more like agents,

leading to positive impacts in some situations and negative impacts in others). This therefore tests

for whether the influences on managers help explain a positive versus negative impact of discretion

on performance and thereby potentially help explain the contradictory empirical results that give

rise to the discretion puzzle.


of analysis and modelling methodologies have often not sufficiently differentiated

the impact of discretion between different influences on the unit of analysis.

Differentiating discretion’s impact on performance in a more granular way

(i.e. by characteristics of the unit of analysis such as firm type or firm size) might

then lead to more consistent empirical results across future studies on the impact of

managerial discretion on performance.

Box 2.6: Potential Moderating Effects of Firm Type and Firm Size

This Box explains the rationale for the postulate that managers may use their

discretion differently depending on influences on managers in terms of firm

type and firm size. The reasoning is that (as derived in Sects. 2.3.2 and 2.3.3

based on principal-agent theory and stewardship theory) the manager’s use of

discretion may depend on a number of influences, and these influences in turn

are likely to differ between firm types and firm sizes:

• Firm type (i.e. domestic Chinese firms versus foreign multinationals).There are various differences between firm types such as domestic Chinese

firms and foreign multinationals that might affect a manager’s behaviour

and therefore the way in which discretion affects performance. For exam-

ple, stewardship theorists have argued that differences in an organisation’s

situational characteristics (i.e. management philosophy and cultural

differences) can influence whether managers use discretion to the benefit

or at the expense of performance (see Sect. 2.3.3.2; e.g. Davis et al. 1997b,

pp. 32–38; Lawler 1986, 1992; Van Slyke 2007, p. 164; Walton 1980,

1985). In particular, as to cultural differences, individualism and high

power distance are postulated to make the assumptions of principal-

agent theory more likely (which imply that discretion reduces perfor-

mance) whereas collectivism and low power distance make the

assumptions of stewardship theory more probable (which imply that dis-

cretion improves performance; e.g. Davis et al. 1997b, p. 43; Triandis

1990, 1995; Triandis et al. 1993; Vargas Sanchez 2004). These cultural

differences are closely linked to firm type, e.g. with firms in the United

States tending to exhibit high individualism and low power distance and

firms in Japan low individualism and high power distance (e.g. Crossland

and Hambrick 2007, p. 14; Davis et al. 1997b, p. 42; Hofstede 1980, 1991).

Likewise, domestic Chinese firms and foreign multinationals may differ in

terms of both the extent of individualism and the degree of power distance,

with rigid hierarchies of Chinese firms impeding team problem-solving

and joint decision-making (e.g. Aminpour and Woetzel 2006, pp. 46–47;

Hanne Chen 2004, p. 128; Crossland and Hambrick 2007, p. 14;

Hexter 2006, pp. 4–6; Hexter and Woetzel 2007a, p. 4; Hoover 2006,(continued)


p. 92; McGregor 2005, pp. 273–274).102 Hence, for such reasons as

differences in management philosophy and culture, managers working in

Chinese firms might respond differently to additional discretion than

managers in multinationals. This motivates the study to test whether firm

type moderates discretion’s performance impact (Hypothesis 2).

• Firm size (e.g. 150 versus 5,000 employees). By virtue of similar

reasoning, one might expect a manager in a smaller firm (e.g. 150

employees) to respond differently to additional discretion than a manager

in a larger firm (e.g. 5,000 employees). As with firm type, there are various

indications in the literature that a firm’s size might affect managerial

behaviour. For instance, firms of different sizes may differ in their moni-

toring control mechanisms, which aim to reduce the principal’s asymmetry

of information regarding the manager’s actions and thereby aim to

improve the impact of discretion on performance in principal-agent theory

(see Sect. 2.3.2; e.g. Berger et al. 1997, p. 1411; Burkart et al. 1997, p. 705;

Caza 2007; 2011; Chang and Wong 2003, p. 6; Eisenhardt 1989, p. 60;

Fama and Jensen 1983b; Levinthal 1988, p. 153; Spremann 1987,

pp. 10–11; Verhoest 2003, pp. 2–5; Wang et al. 2008; Werner and Tosi

1995, p. 1673). Specifically, Caza (2007, pp. 12–20) argues that a larger

firm with multiple units which are similar to the one led by the manager

under consideration provides the principal with more informational

benchmarks to assess the manager’s performance and therefore potentially

constrain the manager’s misconduct.103 With the influences on the man-

ager as specified in principal-agent theory and stewardship theory thus

potentially differing by firm size, the present study postulates that the

102 Similarly, with different types of firms attracting different people, the psychological

characteristics and expectations that stewardship theorists have argued to influence how managers

use discretion might differ between Chinese firms and multinationals (see Sect. 2.3.3.2; e.g.

Argyris 1973a, b; Brown 1969; Caza 2007; Davis et al. 1997b, pp. 38–43; French and Raven

1959; Gibson et al. 1991; Katz and Kahn 1978; Khanchel 2009, p. 98; Maslow 1970;

McGregor 1960; Mills and Keast 2009; Simon 1957a, b; Turner 1981; Van Slyke 2007;

Vargas Sanchez 2004, pp. 4–5, 2005, pp. 24–25). In addition to these influences derived from

stewardship theory that can affect the manager’s natural predisposition to performance

maximisation, Chinese firms and multinationals might also differ in terms of their compensation

control mechanisms, which can moderate the impact of discretion on performance in principal-

agent theory (see Sect. 2.3.2; e.g. Agrawal and Knoeber 1996, p. 378; Berger et al. 1997, p. 1411;

Burkart et al. 1997, p. 705; Chang and Wong 2003, p. 6; Cheng et al. 2006; Eisenhardt 1989, p. 60;

Fama and Jensen 1983a, p. 345; Jensen and Meckling 1976; Jensen and Murphy 1990, p. 226;

Levinthal 1988, p. 153; Spremann 1987, p. 10; Wang et al. 2008; Werner and Tosi 1995, p. 1673;

Zhang and Li 2008b).103 For the unit of analysis of the present study (see Chap. 3), this would predict that top

management at corporate headquarters in China could, ceteris paribus, better monitor the plant

manager in a larger firm with multiple plants and plant managers than in a smaller firm with only a

single plant and a single plant manager.


impact of managerial discretion on performance potentially differs by firm

size (which is tested in terms of Hypothesis 3; see Box 2.5).

In sum, managers in firms of different types and different sizes might be

subject to different influences specified by principal-agent theory and stew-

ardship theory (see Sects. 2.3.2 and 2.3.3): It has been explained that managers

in Chinese firms might have different interests in maximising performance

than managers in multinationals, and that managers in small firms might be

subject to different monitoring than managers in large firms. With the

manager’s interests in maximising performance and monitoring control

mechanisms affecting how managers use their discretion in principal-agent

theory and stewardship theory, it is thus postulated that the impact of discretion

on performance differs by organisational contexts, i.e. by firm type and firm

size. Whether or not firm type and firm size moderate the impact of discretion

on performance in practice is tested by Hypotheses 2 and 3.

It should be noted that this study’s new discretion model (see Chap. 4) for

empirically testing Hypotheses 2 and 3 simultaneously differentiates the

impact of discretion on performance by firm type and firm size (as well as by

the dimensions of discretion). This allows testing for three-way interaction

(Henseler and Fassott 2010, p. 722), whereby firm type moderates the

moderating effects of firm size on the impact of discretion on performance.

In simple terms, the model allows firm size to differently influence how a

manager uses discretion according to whether the firm type is a Chinese firm or

a multinational. The need to do so derives from the analysis of principal-agent

theory in Sect. 2.3.2.2 (see Box 2.2), which has demonstrated that one moder-

ator variable (e.g. the manager’s interests in maximising performance; related

to firm type above) can moderate the moderating effects of other moderator

variables (e.g. monitoring control mechanisms; related to firm size above).104

Finally, as emphasised throughout this chapter, it is reiterated that while firm

type and firm size provide examples of potential moderator variables that may

influence theway that discretion is used to impact on performance, there are also

a number of other moderators that could be investigated in future research.

For example, onemight attempt tomodel the aforementioned influences derived

from principal-agent theory and stewardship theory directly rather than

modelling differences in firm type and firm size that may encompass several(continued)

104 For example, if plant managers in Chinese firms tended to be more performance-maximising

than top management but plant managers in multinationals tended to be less performance-

maximising than top management, then better monitoring control mechanisms in e.g. larger

firms that aligned the plant manager’s actions more closely with those desired by top management

would reduce performance in Chinese firms but improve performance in multinationals. Firm type

would then influence (i.e. moderate) whether firm size has a positive or negative moderating effect

on the impact of discretion on performance. This study’s model allows for such three-way

interaction by estimating separate moderating effects of firm size by firm type.


of these moderators at once.105 For the present study’s research objective,

however, it is sufficient to take firm type and firm size as examples of greater

granularity, since finding that the impact of discretion on performance differs by

firm type and/or firm size would constitute a proof-by-counter-example that

greater granularity can be important (see Sect. 1.2). Choosing firm type and firm

size as the particular examples for demonstrating granularity’s importance is

appealing for a number of reasons, including that (1) studies finding positive,

neutral, and negative impacts of discretion on performance have often differed

in terms of firm type and firm size (see Sect. 2.2; e.g. Adams et al. 2005;

Huiyuan Chen 2006; Cheng et al. 2006; Groves et al. 1994; Heinecke 2011;

Lopez-Navarro and Camison-Zornoza 2003; Thomas and Peyrefitte 1996;

Wang et al. 2008; Werner and Tosi 1995; Williamson 1963; Xu et al. 2005;

Yan et al. 2010; Zhang and Li 2008b; Zhang and Xie 2008; Zhang 1997),

(2) empirical studies have rarely tested for whether differences in firm type

and firm size might moderate the impact of discretion on performance, and

(3) unlike the largely psychology-oriented moderators proposed by the existing

theories, firm type and firm size lend themselves well to both empirical verifica-

tion and practical application.106

The common theme of Hypotheses 1, 2, and 3 is the present study’s postulate

that greater granularity than in existing theories and many empirical studies will

yield a better explanation of the impact of discretion on performance and thereby

contribute to the resolution of the discretion puzzle (see Box 2.5): Managers may

105 Empirical evidence exists for both approaches. On the one hand, scholars have tested

moderators that derive directly from the assumptions of principal-agent theory and stewardship

theory, such as managerial incentives (Cheng et al. 2006), managerial commitment, the number of

similar units, managerial education, and managerial experience (Caza 2007; 2011), top manage-

ment pay gap (Zhang and Li 2008b), and corporate control via performance monitoring, incentive

systems, and social integration (Wang et al. 2008). On the other hand, scholars have tested

moderators that aggregate individual influences to an organisational context, such as ownership

concentration (Yougen Li and Zhao 2004), export joint venture group composition

(Lopez-Navarro and Camison-Zornoza 2003), the organisational type of state-owned enterprises

(Xu et al. 2005), and market competition (Zhao et al. 2010). The organisational context measured

by these moderators may tap into individual influences from the theories, e.g. competition might

tap into monitoring control mechanisms: ‘The firm is disciplined by competition from other firms,

which forces the evolution of devices for efficiently monitoring the performance of the entire team

and of its individual members’ (Fama 1980, p. 289).106 As noted in Sect. 2.3.3.2, even if it were empirically proven that certain psychological

characteristics, perceptions, and expectations predicted whether a given manager acted as a

steward versus an agent, it would be intricate for a principal in practice to observe these factors

and thus evaluate ex ante whether or not the manager would act as a steward or an agent

(Davis et al. 1997b, p. 22; Williamson 1985)—making it difficult to recommend whether to

increase or decrease managerial discretion in practice.


use managerial discretion differently depending on the area of their work in which

discretion is granted (i.e. Hypothesis 1: dimension of discretion) and the influences

on the managers in place (i.e. Hypothesis 2: firm type; Hypothesis 3: firm size). The

impact of discretion on performance could then differ by the dimensions of

discretion (e.g. capital investment discretion versus hiring discretion), firm type

(e.g. Chinese firms versus multinationals), and firm size (e.g. 150 versus 5,000

employees)—and this impact could empirically be estimated as positive, neutral or

negative depending simply on how discretion is measured (i.e. dimensions of

discretion) and how the unit of analysis is differentiated (i.e. firm type and firm

size; see Sect. 7.1).

If the present study’s postulate were to be confirmed, one potential explanation

towards resolving the discretion puzzle would be that existing research has often

fallen short of differentiating the impact of discretion by discretion dimensions and

influences on managers and that this limited granularity could be responsible for the

contradictory results for the overall effect of discretion on performance (e.g.

Agarwal et al. 2009; Barnabas and Mekoth 2010; Bowen et al. 2008; Caza 2007;

2011; Chang and Wong 2003; Huiyuan Chen 2006; Cheng et al. 2006;

Gammelgaard et al. 2010; He et al. 2009; Khanchel 2009; Yougen Li and Zhao

2004; Stano 1976; Werner and Tosi 1995; Williamson 1963; Zhang and Li 2008b;

Zhang 1997). This would show by virtue of proof-by-counter-example that discre-

tion, instead of universally having a single positive, neutral or negative effect on

performance, can in fact increase, not alter or reduce performance depending on the

area of the manager’s work in which discretion is granted (i.e. dimension of

discretion) and the influences on managers in place (i.e. firm type and firm size).

Future empirical and theoretical research would then need to differentiate the

impact of discretion by this greater granularity (e.g. in terms of dimensions, firm

type, and firm size) in order to obtain consistent rather than contradictory results

across studies. Empirical studies might then find that given certain influences on

managers, certain dimensions of discretion (e.g. discretion for capital investments)

tended to consistently improve performance whereas other dimensions tended to

reduce performance. Furthermore, the present study might then be able to provide

guidance as concerns the direction of future research on advancing theories in an

effort to resolve the discretion puzzle, which is discussed in the next section on

Hypothesis 4.

2.4.3 Implications for Postulate and Hypotheses: Hypothesis 4

In addition to allowing the present study to formulate Hypotheses 1 to 3 (see

Sect. 2.4.2), the thorough and up-to-date review of the empirical and theoretical

literature permits this study to derive its final hypothesis, Hypothesis 4. Whereas

Hypotheses 1 to 3 test whether greater granularity is important for resolving the

discretion puzzle, Hypothesis 4 tests whether the existing theories can fully explain

the empirical evidence that results when adopting this greater granularity. Hypoth-

esis 4 therefore tests the applicability of those theories for investigating the impact


of managerial discretion on performance after taking into account the required

greater granularity.

Hypothesis 4 is derived by integrating the literature on the impact of discretion on

performance across the columns in the literature review summary Table 2.12 above.

Referring to the columns in Table 2.12 with respect to the row ‘Impact on performance:

Direct’, ‘Empirical Evidence’ provides contradictory results ranging from positive to

neutral and even negative estimated impacts of discretion on performance (see

Sect. 2.2; e.g. Agarwal et al. 2009; Barnabas and Mekoth 2010; Caza 2011; Chang and

Wong 2003; Gammelgaard et al. 2010; Groves et al. 1994; He et al. 2009;

Heinecke 2011; Khanchel 2009; Yougen Li and Zhao 2004; Lopez-Navarro and

Camison-Zornoza 2003; Stano 1976; Venaik 1999; Williamson 1963; Xu et al.

2005). As explained in Sect. 2.4.1, this gives rise to the discretion puzzle (see

Sect. 1.1), because the existing theories cannot fully explain this contradictory

evidence—with ‘Managerial Discretion Theory’ not specifying the impact of discretion

on performance (see Sect. 2.3.1), ‘Principal-Agent Theory’ predicting mostly a nega-

tive impact (see Sect. 2.3.2), and ‘Stewardship Theory’ predicting a positive impact of

discretion on performance (see Sect. 2.3.3).

The discretion puzzle is depicted in the fictitious scatter plot in Fig. 2.6, which

integrates the fictitious scatter plots from Figs. 2.1, 2.2, 2.3, 2.4 and 2.5 in this chapter.

With the horizontal axis measuring the extent of the manager’s discretion (D) and thevertical axis the resulting performance (P ), the upwards-sloping line represents a

positive, the horizontal line a neutral (or insignificant), and the downwards-sloping



HighManagerial

Discretion (D)

Low Performance (P)

DiscretionPuzzle

Neutral (d ≈ 0) (i.e. null hypothesis for Hypothesis 4)

(e.g. Caza, 2011; Groves, et al., 1994;Y. Li & Zhao, 2004; López-Navarro &Camisón-Zornoza, 2003; Venaik,1999)

Fig. 2.6 Fictitious scatter plot for Hypothesis 4

Source: Selection of empirical studies on the impact of managerial discretion on performance;

principal-agent theory and stewardship theory


line a negative impact of managerial discretion on performance.107 The contradictory

(i.e. positive, neutral, and negative) empirical evidence on the impact of managerial

discretion on performance that gives rise to the discretion puzzle is depicted by

simultaneously having upwards-sloping, horizontal, and downwards-sloping lines.

Furthermore, Fig. 2.6 integrates the opposing predictions of the existing theories on

the impact of discretion on performance, which are revisited briefly below with

reference to this study’s unit of analysis (i.e. the plant manager in China):

• Managerial discretion theory (see Sect. 2.3.1; e.g. Adams et al. 2005; Crossland and

Hambrick 2007; Finkelstein and Hambrick 1990; Hambrick and Finkelstein 1987;

Misangyi 2002; Quigley and Hambrick 2009; Tang 2008; Zhao et al. 2010) implies

that extending a plant manager’s discretion for making capital investments from say

zero RMB (left in Fig. 2.6) to 100,000 RMB (right in Fig. 2.6) gives the plant

manager the opportunity to make investments in a way that might either positively

or negatively affect performance. Increasing discretion (moving rightwards on

Fig. 2.6) thus grants the plant manager greater latitude of action to influence

performance—so performance is expected to deviate more strongly from average

performance (the horizontal axis) as discretion rises—but it is left unspecified

whether this impact is positive (upwards-sloping) or negative (downwards-sloping).

• Principal-agent theory (see Sect. 2.3.2; e.g. Caza 2007, p. 10; Caza 2011;

Chang and Wong 2003, p. 7; Davis et al. 1997b, p. 38; Hutzschenreuter and

Kleindienst 2007, p. 4; Jensen and Murphy 1990; Zhao et al. 2010) predicts that

granting the plant manager additional discretion (moving rightwards on Fig. 2.6)

will mostly reduce performance (downwards-sloping line in Fig. 2.6), since the

plant manager maximises utility by pursuing his/her self-serving interests

(assumption 1) and given that top management at corporate headquarters in

China cannot fully monitor the plant manager’s actions (assumption 2), he/she

will choose to undertake such capital investments that produce the plant

manager’s desired outcomes (assumption 3) but tend to come at the cost of the

principal and reduce performance (assumption 4).

• Stewardship theory (see Sect. 2.3.3; e.g. Corbetta and Salvato 2004; Davis et al.

1997b, pp. 25–26; Donaldson and Davis 1991, p. 52; Eddleston and Kellermanns

2007, p. 547; Khanchel 2009, pp. 98–99; Liu and Cai 2004, p. 4; Mills and Keast

2009, pp. 14–15; Tosi et al. 2003, p. 2054; Van Slyke 2007, pp. 165–167;

Vargas Sanchez 2004, p. 3, 2005, pp. 18–19) postulates that extending the plant

manager’s discretion (moving rightwards on Fig. 2.6) will improve performance

(upwards-sloping line in Fig. 2.6), because the plant manager maximises utility by

achieving organisational instead of self-serving objectives (assumption 1), so

although top management cannot fully monitor the plant manager’s actions

107 In terms of the algebraic notation utilised throughout this chapter, Fig. 2.6 depicts performance as

a linear function of discretion (i.e.P ¼ d � Dþ c � Cþ m � D �Mð Þ), with the impact of discretion on

performance (i.e. @P @D= ¼ d þ m �M) represented by the slopes of the lines. When the impact of

discretion on performance is positive (@P @D= > 0), the line is upwards-sloping and when the impact

is negative, it is downwards-sloping.


(assumption 2), he/she will diligently choose such actions that produce the plant

manager’s desired outcomes (assumption 3), namely enhancing performance

(assumption 4).

Figure 2.6 therefore reveals that the existing theories do not unequivocally

explain the contradictory empirical evidence on the impact of discretion on perfor-

mance in the literature (see Sect. 1.1). As managerial discretion theory leaves the

impact of discretion on performance unspecified while stewardship theory specifies

this impact as positive (upwards-sloping line) and principal-agent theory as mostly

negative (downwards-sloping line), no single theory can fully explain the coexis-

tence of positive, neutral, and negative impacts of discretion on performance in the

empirical literature.108 As the existing empirical evidence is, however, mainly

based on limited granularity (see Sects. 1.1 and 2.2.3; e.g. Bowen et al. 2008;

Chang and Wong 2003; Gammelgaard et al. 2010; Groves et al. 1994; He et al.

2009; Xiaoyang Li 2007; Stano 1976; Venaik 1999; Werner and Tosi 1995; Zhang

1997), it remains to be seen whether the existing theories can adequately explain the

impact of discretion on performance after applying the greater granularity that is

hypothesised to help resolve the discretion puzzle in the present study (i.e.

Hypotheses 1 to 3).

This is empirically verifiedwithin the present study by testing Hypothesis 4 (see Box

2.7), which tests the postulate that the empirical evidence will not always be consistent

with any one of the existing theories even after differentiating the impact of managerial

discretion on performance by the dimensions of discretion (Hypothesis 1), firm type

(Hypothesis 2), and firm size (Hypothesis 3). Specifically, the postulate is verified by

testing the null hypothesis (H04) that discretion does not alter performance (i.e. horizontal

line in Fig. 2.6; which is predicted by neither principal-agent theory nor stewardship

theory)109 against the alternative hypothesis from principal-agent theory (H14 ) that

discretion decreases performance (i.e. downwards-sloping line) and the alternative

108 As discussed in Sects. 1.1 and 2.3, existing combinations of managerial discretion theory,

principal-agent theory, and stewardship theory also do not fully explain the impact of discretion on

performance (e.g. Caza 2007; 2011; Davis et al. 1997b, pp. 27–43; Finkelstein and Peteraf 2007,

pp. 237–243; Lane et al. 1999, p. 1079; Vargas Sanchez 2001, 2004, 2005).109 As explained in Sect. 2.3.1, managerial discretion theory does not clearly specify the nature of

the impact of discretion on performance and therefore makes no predictions on e.g. positive versus

negative impacts (e.g. Adams et al. 2005; Crossland and Hambrick 2007; Finkelstein and

Hambrick 1990; Hambrick and Finkelstein 1987; Misangyi 2002; Quigley and Hambrick 2009;

Tang 2008; Zhao et al. 2010). Although principal-agent theory can be consistent with the null

hypothesis (H04) that discretion does not alter performance in exceptional cases—i.e. when control

mechanisms are so effective that they unhinge the theory’s assumptions (see Box 2.1 and Box 2.2

in Sect. 2.3.2.2; e.g. Agrawal and Knoeber 1996, p. 377; Chang and Wong 2003, p. 7)—the main

prediction of principal-agent theory is the alternative hypothesis (H14 ) that discretion decreases

performance (see Sect. 2.3.2; e.g. Caza 2007, p. 10; Caza 2011; Chang and Wong 2003, p. 7;

Davis et al. 1997b, p. 38; Hutzschenreuter and Kleindienst 2007, p. 4; Jensen and Murphy 1990;

Zhao et al. 2010).


hypothesis from stewardship theory (H24 ) that discretion increases performance (i.e.

upwards-sloping line):

Box 2.7: Postulate and Hypothesis 4

Postulate of the Present StudyEmpirical evidence will not always be consistent with any one of the existing

theories after applying the greater granularity that is hypothesised to help

resolve the discretion puzzle in the present study.

Hypothesis 4 (Applicability of Principal-Agent Theory and StewardshipTheory)

H04:The impact of managerial discretion on performance is zero (d ¼ 0).110

H14:The impact of managerial discretion on performance is negative (d < 0;

principal-agent theory).

H24 :The impact of managerial discretion on performance is positive (d > 0;

stewardship theory).

• If the null hypothesis (H04) can be universally rejected in favour of the principal-

agent hypothesis (H14),

111 then principal-agent theory can explain the empirical

results in the present study after accounting for the greater granularity. Likewise,

if the null hypothesis (H04) can be universally rejected in favour of the steward-

ship hypothesis (H24 ), then stewardship theory can unequivocally explain the

empirical results in the present study.

• However, if results are mixed, e.g. capital investment discretion increases

performance (H24 ) but hiring discretion decreases performance (H1

4 ), then the

study’s postulate is confirmed: Neither principal-agent theory nor stewardshiptheory could then explain the impact of discretion on performance after account-

ing for greater granularity.

110 In terms of the algebraic notation utilised throughout this chapter, the parameter dmeasures the

direct effect of managerial discretion (D) on performance (P), which is equal to the total impact of

discretion on performance (@P @D= ¼ d þ m �M) when moderators are zero (M ¼ 0), i.e. when

they take their average values (see Sect. 4.3.2 on comparative statics; e.g. Aiken and West 1991,

p. 37; Dowling 2000, pp. 284–291; Finney et al. 1984; Henseler and Fassott 2010, p. 728;

Hirschey 2009, p. 99).111 Universally rejected refers to rejecting the null hypothesis after accounting for the greater

granularity, e.g. rejecting the null hypothesis when differentiating discretion’s impact by multiple

dimensions of discretion.


• Moreover, such mixed results would provide evidence that the existing reconcil-iation of principal-agent theory and stewardship theory, which postulates that a

manager is either a steward or an agent (see Sect. 2.3.3.2; e.g. Davis et al. 1997b,

pp. 27–43; Lane et al. 1999, p. 1079; Vargas Sanchez 2001, 2004, 2005), is not

empirically applicable (see Sect. 7.4.2). Instead of acting as a steward (using

discretion to the benefit of performance) or as an agent (using discretion at the

expense of performance), such mixed results would indicate that managers can

exhibit mixed behaviour that is not accounted for by the existing theories—such

as acting as a steward when making capital investments and acting as an agent

when hiring workers.

By testing Hypothesis 4, the present study therefore not only provides evidence

on the applicability of principal-agent theory and stewardship theory for the chosen

unit of analysis (i.e. middle management in China), but also works towards resolv-

ing the discretion puzzle by highlighting the potential insufficiency of the existing

theories by virtue of proof-by-counter-example.112 Evidence that the theories are

individually insufficient would vindicate the present study’s new discretion model

(see Chap. 4) in its approach to combine managerial discretion theory, principal-

agent theory, and stewardship theory and furthermore would provide a new starting

point for developing theories in future research on managerial discretion so as to

eventually resolve the discretion puzzle (see Sect. 7.4.3).

2.4.4 Implications for Unit of Analysis

In addition to yielding the research gap and research objective (see Sect. 2.4.1) as

well as the postulate and hypotheses (see Sects. 2.4.2 and 2.4.3), the present study’s

unit of analysis can be selected based on the review of the literature summarised in

Table 2.12 above (see the row ‘Unit of analysis’):

• The empirical literature (see the column ‘Empirical Evidence’ in Table 2.12)

reviewed in Sect. 2.2 has revealed that while there have been individual studies

for middle management and China, compared to the total amount of studies,

evidence on the impact of discretion on performance has remained particularly

scarce for both middle management (see Caza 2007, p. 1) and for China (see

Yougen Li and Zhao 2003, p. 6; Zhang and Li 2008a, pp. 37–38). It follows that

given the abundance of middle managers in organisations113 and the importance

112 If the empirical results are not consistent with any single existing theory for the present study’s

particular instance of plant managers in China, this would serve as a proof-by-counter-example for

the aforementioned postulate that neither existing theory can fully explain the empirical impact of

discretion on performance after applying the greater granularity, since then neither theory could

fully explain the impact of discretion on performance at least in the particular instance of the

present study and therefore not universally in all studies.113 As explained in Chap. 3, there are tens of thousands of plant managers in China alone

(Guojia tongji ju [National Bureau of Statistics] 2007, 14–1, 14–2, 14–18). With plant managers

being but one example of middle managers, this translates into an even larger number of middle

managers in organisations worldwide.


of China for domestic Chinese firms and foreign multinationals (Aminpour and

Woetzel 2006, p. 41; Grant 2006, p. 25; Hexter 2006, p. 1; Hoover 2006, p. 92;

Kaufmann et al. 2005, p. 21; McGregor 2005, pp. 2, 272; Pascha 1998, p. 57;

Taube 2008, p. 186; Tian 2007, pp. 7–8), this limited evidence on the impact of

discretion motivates the choice of middle management in China as the unit of

analysis.

• The theoretical literature (see the columns ‘Managerial Discretion Theory’,

‘Principal-Agent Theory’, and ‘Stewardship Theory’ in Table 2.12) reviewed

in Sect. 2.3 has vindicated this choice of the unit of analysis by demonstrating

that the three relevant theories are all applicable at the middle management level

(see Caza 2007, pp. 7–8; Davis et al. 1997b, p. 25; Jensen and Meckling 1976,

p. 309). Following the initial focus of managerial discretion theory (see

Sect. 2.3.1), principal-agent theory (see Sect. 2.3.2), and stewardship theory

(see Sect. 2.3.3) on top management, the theories have been extended to a

broad range of alternative units of analysis, including middle management

(e.g. Acemoglu et al. 2007; Arthurs and Busenitz 2003; Barnabas and Mekoth

2010; Baysinger and Butler 1985; Berger et al. 1997; Bloom et al. 2008;

Bogart 1995; Brody 1996; Brush et al. 2000; Caza 2007; 2011; Chang and

Wong 2003; Childs and Mauer 2008; Corbetta and Salvato 2004; Denis et al.

1997; Dharwadkar et al. 2000; Dicke and Ott 2002; Donaldson and Davis 1989,

1991, 1994; Eddleston and Kellermanns 2007; Fox and Hamilton 1994;

Glaister et al. 2003; Khanchel 2009; Lee and O’Neill 2003; Lopez-Navarro

and Camison-Zornoza 2003; McGubbins et al. 1987; Miller and Le Breton-

Miller 2006; Mills and Keast 2009; Muth and Donaldson 1998; Oh 2002;

Ongore 2011; Perrone et al. 2003; Salvato 2002; Tian and Lau 2001;

Tosi et al. 2003; Van Slyke 2007; Vargas Sanchez 2001, 2004; Venaik 1999;

Walters 1995; Wood and Waterman 1991; Zahra 2003).

In short, the scarcity of empirical evidence for middle management and for

China as well as the applicability of managerial discretion theory, principal-agent

theory, and stewardship theory to the middle management level encourage this

study to choose middle management in China as its unit of analysis (as noted in

Table 2.12 above). Specifically, the present study’s unit of analysis is the plant

manager of small and medium-sized enterprises in the manufacturing sector

throughout mainland China in the latter half of 2007, which is discussed in detail

in the next chapter (see Chap. 3).

2.4.5 Implications for Model

Finally, the completed review of the literature summarised in Table 2.12 above

allows Chap. 4 to develop a new empirically-validated model on the impact of

managerial discretion on performance. As described in Chap. 4, the new discretion

model integrates the empirical literature and the theoretical literature in a way that

permits this study to test its hypotheses for the chosen unit of analysis so as to fulfil

the research objective of working towards resolving the discretion puzzle:


• The theoretical literature on managerial discretion theory, principal-agent the-

ory, and stewardship theory (see Sects. 2.1 and 2.3) is synthesised into the

study’s new discretion model in Chap. 4. In line with the contributions of

the theories highlighted in Table 2.12 above, the new model integrates the

theoretical context specified by managerial discretion theory (e.g. regarding

the potentially multidimensional construct and the antecedents of discretion)

and the theoretical mechanisms specified by principal-agent theory and steward-

ship theory (e.g. direct and moderating effects). In this way, the present study

synthesises the existing theoretical content of the three relevant theories and

their extant combinations into a new theoretical discretion model.• The empirical literature (see Sects. 2.1 and 2.2) is likewise harnessed for

developing the present study’s new discretion model in Chap. 4. Since the

review of the existing empirical evidence in Sect. 2.2 has revealed that the

impact of discretion on performance might need to be investigated with greater

granularity so as to resolve the discretion puzzle, the theories are woven into the

new model with more fine-grained granularity than in many previous studies.

In particular, the model is specified so as to differentiate the impact of discretion

on performance by the granularity required for testing the study’s hypotheses

(i.e. simultaneously differentiating by dimension of discretion, firm type, and

firm size). Moreover, following the synthesis of the existing theories into the

theoretical discretion model, state-of-the-art modelling methodology from the

literature is utilised to translate the theoretical discretion model into the new

empirical discretion model, which is calibrated within the present study.

By combining the available body of empirical and theoretical knowledge

reviewed in this chapter, the present study thus develops a new empirically-

validated discretion model for investigating the impact of managerial discretion

on performance. This model synthesises managerial discretion theory, principal-

agent theory, and stewardship theory while taking the greater granularity from the

empirical literature into account in a way that allows the present study to test its

hypotheses and thereby fulfil its research objective of working towards resolving

the discretion puzzle.

In sum, this chapter has conducted a thorough, in-depth review of both the

empirical and the theoretical literature on the impact of managerial discretion on

performance. It was shown in this section that this literature review summarised in

Table 2.12 above enables the present study to derive its research gap and research

objective (see Sect. 2.4.1), formulate its postulate and hypotheses (see Sects. 2.4.2

and 2.4.3), select a unit of analysis (see Sect. 2.4.4), and develop a new discretion

model (see Sect. 2.4.5). With the research gap and research objective defined in

detail in Sects. 1.1 and 1.2, and the postulate and hypotheses fully specified in this

section, the next two chapters scrutinise the unit of analysis (see Chap. 3) and the

study’s new discretion model (see Chap. 4). Once the new discretion model has

been calibrated and empirically-validated (see Chap. 5), it will become possible to

utilise the model to test the present study’s four hypotheses (see Chap. 6) and

thereby generate far-reaching conclusions that work towards resolving the discre-

tion puzzle—i.e. that fulfil the research objective (see Chap. 7).


Documents

Managerial Discretion and Performance in China - ReadingSample