"Tying the Manager's Hands": How Firms can make Credible Commitments that make Opportunistic Managerial Intervention Less Likely

Abstract

We discuss and empirically examine a firm-level equivalent of the ancient problem of "tying the King’s hands", namely how to maximize managerial intervention for "good cause", while avoiding intervention for "bad cause". Managers may opportunistically intervene when such intervention produces private benefits. Overall firm performance is harmed as a result, because opportunistic managerial intervention harms employee motivation. The central point of the paper is that various mechanisms and factors, such as managers staking their personal reputation, employees controlling important assets, strong trade unions, corporate culture, etc. may function as constraints on managerial proclivities to opportunistically intervene. Thus, firms can make credible commitments that check managerial proclivities to opportunistically intervene. We derive 5 hypotheses from these ideas, and test them, using path-analysis, on a rich dataset, based on 329 firms in the Spanish food and electric/electronic industries.

Full text

DANISH RESEARCH UNIT FOR INDUSTRIAL DYNAMICS
DRUID Working Paper No 03-10
“Tying the Manager’s Hands”:
How Firms Can Make Credible
Commitments That Make Opportunistic
Managerial Intervention Less Likely
by
Kirsten Foss, Nicolai J. Foss and Xosé H. Vàzquez-Vicente

“Tying the Manager’s Hands”: How Firms Can Make Credible
Commitments That Make Opportunistic Managerial
Intervention Less Likely
Kirsten Foss
Department of Industrial Economics and Strategy
Copenhagen Business School
Solbjergvej 3, 3rd floor; 2000 Frederiksberg
Denmark; kf.ivs@cbs.dk
Nicolai J. Foss
Department of Management, Politics, and Philosophy
Copenhagen Business School,
Blaagaardsgade 23B, DK-2200 Copenhagen N
Denmark njf.lpf@cbs.dk
Xosé H. Vázquez-Vicente
Department of Business Management and Marketing
Faculty of Economics and Business, Universidade de Vigo
Campus das Lagoas-Marcosende, 36200- Vigo
Spain; xhvv@uvigo.es
29 May 2002; 24 September 2002; 10 December 2002; 1 April 2003; 20 August 2003
Abstract
We discuss and empirically examine a firm-level equivalent of the ancient problem of “tying the
King’s hands,” namely how to maximize managerial intervention for “good cause,” while
avoiding intervention for “bad cause.” Managers may opportunistically intervene when such
intervention produces private benefits. Overall firm performance is harmed as a result, because
opportunistic managerial intervention harms employee motivation. The central point of the
paper is that various mechanisms and factors, such as managers staking their personal reputation,
employees controlling important assets, strong trade unions, corporate culture, etc. may function
as constraints on managerial proclivities to opportunistically intervene. Thus, firms can make
credible commitments that check managerial proclivities to opportunistically intervene. We
derive 5 hypotheses from these ideas, and test them, using path-analysis, on a rich dataset, based
on 329 firms in the Spanish food and electric/electronic industries.
Key words: Managerial opportunism, credible commitments, organizational design, transaction
cost economics
JEL Codes: D23, D74, L22, M12,
ISBN 87-7873-142-9

Acknowledgements
The comments of Sven Haugland, Thorbjørn Knudsen, Torben Pedersen, Frank Stephen and
Xosé M. García Vázquez are gratefully acknowledged. The field survey has benefited from
financial support by SXID, a Research Unit of the Galician Government, through grant
PROY99–10, and by CICYT, an agency of the Spanish Government through grant SEC99-
1191.

1
I. Introduction
In this paper we discuss and empirically examine a firm-level equivalent of the ancient problem
of “tying the King’s hands” (Root 1989). A key theme in much of the work on the theory of the
firm (e.g., Coase 1937; Malmgren 1961; Casson 1994; Williamson 1996; Foss 1997; Wernerfelt
1997) is that the exercise of authority in the form of managerial fiat in response to changes in the
environment provides a reason why firms exist. Such managerial intervention will typically
override existing instructions of employees, and in firms where employees are given considerable
discretion, managerial intervention may furthermore amount to overruling decisions made by
these employees on the basis of delegated decision rights.
A fundamental (though arguably somewhat neglected) set of problems is that the option to
intervene (1) “… can be exercised both for good cause (to support expected net gains) and for bad
(to support the subgoals of the intervenor)” (Williamson 1996: 150-151), (2) it may be difficult to
verify the nature of the cause, and (3) promises to only intervene for good cause are hard to make
credible because they are not enforceable in a court of law. There is thus a problem of “…
credibly [promising] to respect autonomy save for those cases where expected net gains to
intervention can be projected” (Williamson 1993: 104). A primary challenge  in theory as
well as managerial practice  is therefore how to maximize managerial intervention for “good
cause,” while avoiding intervention for “bad cause.”1
In this paper, we contribute to the understanding of this problem by examining how firms
can make credible commitments that make (perceived and/or real) opportunistic managerial
intervention (Dow 1987; Kreps 1990), “intervention for bad cause,” less likely. Our overarching
perspective on these issues is mainly drawn from organizational economics (e.g., Milgrom 1988;
Jensen and Meckling 1992; Bijl 1996; Milgrom and Roberts 1996; Williamson 1996; Aghion and
Tirole 1997; Baker, Gibbons, and Murphy 2000), and on political economy work on credible
commitments (e.g., Weingast and Marshall 1988; Miller 1992; Miller and Hammond 1994; Moe
1997). However, in order to lend further support for our arguments, we also draw on ideas about
psychological contracts in organizations (Argyris 1960; Rousseau 1989; Coyle-Shapiro and
1 Milgrom and Roberts (1996: 168) argue that “… the very existence of centralized authority is incompatible with a
thorough going policy of efficient selective intervention. The authority to intervene inevitably implies the authority to
intervene inefficiently.” While we agree that “first-best intervention” is strictly impossible, “second-best
intervention” is feasible.

2
Kessler 2000; Tepper and Taylor 2003), extrinsic and intrinsic motivation (Osterloh and Frey
2000), and psychological research on decision-making (e.g., Bazerman 1994).
Our argument begins from the observation that all firms that are larger than the one-man
firm rely on both the use of managerial authority and employee discretion, that is, the ability of
employees to control resources including their own human capital. While authority is needed,
for example, to manage residual interdependencies, discretion may be rationally delegated to
employees, because it stimulates motivation and fosters local learning and the use of local
knowledge. A considerable body of work in organization theory, including organizational
economics, has addressed issues that relate to the distinction between authority and delegation,
such as the optimal span of control (Williamson 1970), the design of information structures
(Galbraith 1974), and optimal delegation given the moral hazard problem (Jensen and Meckling
1992; Armstrong 1994; Aghion and Tirole 1997). In these treatments, authority is a matter of
control and the giving of orders. Other issues that are implied by the distinction between
authority and discretion have arguably been given less attention, notably how the exercise of
authority in the form of opportunistic managerial intervention harm motivation, diminishing the
beneficial effects of discretion.2 From this perspective, a basic problem in organizational design
is that beneficial delegation is hard to sustain under the property rights structure characterizing
the firm in which delegated decision rights are always “loaned, not owned” (Baker et al. 1999).
Thus, those who hold ultimate decision rights (i.e., authority) may use these to renege on
delegation, overrule decisions made on the basis of delegated rights, and selectively intervene for
bsad cause. This harms employee motivation. However, managers may be constrained by
various mechanisms, including implicit contracts (Kreps 1990; Baker et al. 1999) or explicit
credible commitments (Brocker et al. 1992; Moe 1997) that reduce the incidence and severity of
such harmful interventions.
The design of the paper is as follows: We develop a notion of authority that goes beyond
the picking of well-defined actions from an employee’s action set (as in Simon 1951) and also
includes the power to delegate and constrain discretion, as well as the ability to veto
subordinates’ decisions. We also focus on the costs and benefits of delegating discretion to
employees. We then turn to a discussion of the motivational problems that may arise when
managers exercise authority by reneging on the delegation of discretion, that is “opportunistic
2 However, see Rousseau (1989), Robinson and Rousseau (1994), and Robinson and Morrison (1995) for
organizational behavior work, and Aghion and Tirole (1997) and Baker, Gibbons and Murphy (1999, 2002) for
organizational economics work, that has a strong bearing on these issues.

3
managerial intervention.” It is often in an organization’s interest to avoid such managerial
intervention. There are various mechanisms that may credibly constrain the flexibility of
managers to intervene opportunistically. Some of these are external to the firm (e.g., tight labor
and capital markets, strong labor unions), and some are internal to the firm. In the latter category
are credible commitments undertaken by managers themselves (e.g., managers staking personal
reputations), as well as employees controlling critical resources. A number of hypotheses are
derived and tested on data from the Spanish electronics and food industries. To the best of our
knowledge, the present paper represents the first empirical, firm-level work on these issues.
II. Managerial Intervention and Delegation:
Tensions and Credible Managerial Commitment
Authority and Delegation of Discretion
Simon (1951) provides a classic notion of authority. Authority is defined as the situation
in which a “boss” is permitted by a “worker” to select actions, A0 ⊂ A, where A is the set of the
worker’s possible behaviors. For the worker to accept the assignment, it must lie within his
“zone of acceptance.” A limitation of this notion of authority is that it seems to be based on the
boss having all the information, the worker being merely a passive instrument who reacts to
instructions based on this information. This is a notion that does not square easily with the
(alleged) increasing importance of partly self-managing knowledge-workers in modern
production (e.g., Purser 1998).
Simon (1991: 31) himself later noted that authority may be understood more broadly,
namely as a command that takes the form of a result to be produced, a principle to be applied, or
goal constraints, so that “[o]nly the end goal has been supplied by the command, and not the
method of reaching it.” However, even this is arguably too narrow. Usually, some aspects of
“the method of reaching” an end goal are specified, so that employees are seldom granted full
discretion. Indeed, a function of authority is the placing of restrictions on the decision rights that
are granted to employees with respect to how they reach an end goal (Milgrom 1988; Barzel
1997; Holmström 1999). Authority in the sense of placing restrictions on behavior is exercised
in order to avoid costs associated with unwanted externalities, including, but by no means limited
to, the costs of morally hazardous behavior. Such externalities may also include coordination
failures, such as scheduling problems, duplicative efforts (e.g., of market information gathering

4
or R&D), and cannibalization of product markets and other instances of decentralized actions
being inconsistent with the firm’s overall strategic planning. These externalities arise when
employees exercise discretion.
Discretion may be defined as the ability of an agent to exercise control over a resource,
that is, she is able to allocate that resource to a purpose that she, for whatever reason, finds
suitable (Barzel 1997). There are various reasons why firms may delegate discretion. For
example, if the employee is better informed than the manager with respect to how certain tasks
should be carried out, and this knowledge is costly to communicate (Casson 1994; Melumad et
al. 1995), efficient co-location of decision-making rights and knowledge requires that employees
are delegated discretion with respect to how they use their expertise in problem solving (Jensen
and Meckling 1992). Also, delegation may be undertaken for motivational rather than
knowledge-based reasons. Thus, a long tradition in social psychology (probably beginning with
Roethlisberger et al. 1939) and more recently in the empowerment literature (Conger and
Canungo 1988; Thomas and Velthouse 1990; Gal-Or and Amit 1998), suggests that increasing
the delegation of discretion to employees often “… raises the perceived self-determination of
employees and therewith strengthens intrinsic motivation” (Osterloh and Frey 2000: 543). In
turn, this may lead to an increase in creativity in the pursuit of goals.3 Expert knowledge is better
utilized and learning is fostered (Mudambi et al. 2003). In contrast, decreasing the level of
delegated discretion may crowd out intrinsic motivation, particularly when this frustrates the
employee’s “… beliefs regarding the terms and conditions of the reciprocal exchange
agreement” (Rousseau 1989: 23). These arguments suggest the following hypothesis:
Hypotheses 1: Employee motivation depends positively on the degree of delegation of
discretion.
3 Some reservations and potential critiques should be noted at this point. First, it is conceivable that discretion
may harm motivation if employees do not have the knowledge or personality to command such discretion.
Second, employees may feel uncomfortable with increased discretion because it may imply responsibilities
without additional pay or benefits. In short, employees need to have not just the opportunity, but also the ability
and incentive to engage in self-management (cf. Mowday et al. 1982). We hypothesize, however, that on the
aggregate (firm) level, the positive motivational effects of increased delegation dominate the negative ones, and
that opportunity to engage in self-management is at least to some extent matched by a corresponding ability to
do so. Third, in the model we later test, we assume a linear relation between delegation and motivation and
performance. This, too, may be criticized. We have tested whether the inclusion of the squared variable for
delegation improves the goodness of fit and provides a significant coefficient. However, it turns out that the
coefficients are not significant and that, although the absolute and incremental goodness of fit increases slightly,
the parsimounious goodness of fit decreases considerably. We therefore opted for not including this squared
variable.

5
We further argue that the motivational effects of increased delegation give rise to improved
employee productivity. Partial evidence for this is the finding that giving R&D personnel the
right to share research findings with others and to publish such findings increase R&D
productivity (McMillan et al. 2000; Mudambi et al. 2003). In turn, increased employee
productivity causes firm performance to improve.
There are more direct reasons why delegation may improve firm performance. Thus,
employees may be better informed than managers with respect to how certain tasks should be
carried out (Jensen and Meckling 1992). If such knowledge is costly to communicate efficient
co-location of decision-making rights and knowledge then requires that employees are delegated
discretion with respect to how they use their expertise in problem solving (Casson 1994;
Melumad et al. 1995). Furthermore, it is arguable that delegating discretion to employees will
not only lead to a better use of existing knowledge, but also to the discovery of new knowledge
that would not have been discovered in the absence of delegation (Miles et al. 1997). This
reasoning gives rise to the following hypothesis:
Hypotheses 2: Firm performance depends positively on the degree of delegation of
discretion to employees.
So far, nothing has been said about the cost dimensions of delegation. Although we do not test
hypotheses regarding the costs of delegation, we use insights in these costs to develop other
hypotheses. The costs of delegation are treated in the rich agency literature on optimum
delegation (e.g., Jensen and Meckling 1992; Armstrong 1994; Aghion and Tirole 1997; Gal-Or
and Amit 1998). A general conclusion is that delegation creates opportunities for employees to
collect informational rents and/or engage in morally hazardous activities (i.e., “employee
opportunism”). Roughly, optimum delegation obtains when the incremental gain from making
use of expert knowledge equals the incremental costs from loss of control. The cost caused by
control loss is ultimately rooted in the differing preferences of managers and employees in the
relevant hierarchy and the costs of monitoring relevant aspects of the employee’s activities.
Although the agency approach is useful for framing the cost aspects of delegation, it has
certain limitations. First, it builds on an assumption of complete contracting, which makes it
hard to provide a rationale for authority (Hart 1995), except in the limited sense of monitoring.
Second, it abstracts from those costs of delegation that are not the result of moral hazard, but
rather of mistakes, sub-goal optimization, duplicative efforts, wrong timing of decisions and
erroneous co-location of knowledge and decision rights made by entirely well-intentioned

6
employees (cf. Hendry 2002), and which may be reduced by means of the exercise of the
authority.
Thus, in actuality the tradeoff associated with the optimum level of delegation involves
many variables. It may change over time as the relevant determinants change. Given this, a key
management task arguably is to exercise authority in such a way that the organization gropes
towards optimum delegation, and to track the optimum level of delegation in the face of
changing contingencies. Thus, there will necessarily be an interdependence between delegation
and authority. Discussing the issues of authority and discretion separately is, therefore,
problematic because the interdependence between authority and discretion gives rise to distinct
organizational problems, notably the incentive liabilities associated with managerial intervention
that is perceived by employees to be opportunistic.
Managerial Intervention and Changes in Delegation
As discussed earlier, there are both beneficial and negative implications for
organizations of delegating discretion to employees. Thus, firms confront a basic tradeoff in
the choice of delegation. In the absence of managerial intervention, once implemented,
optimum delegation will continue as long as contingencies, such as technology, tastes,
competitive conditions and other external contingencies remain relatively stable, and as long
as managers do not see a need for changing the firm’s overall strategy or the internal resource
allocation in the firm. Given this, there are three overall reasons why optimum delegation
may change, namely, first, as a response to changed external contingencies (Lawrence and
Lorsch 1967; Casson 1994); second, as a result of changed managerial perceptions with
respect to the firms overall strategy and internal resource allocation (Penrose 1959); and,
third, some mix of the two. We briefly consider the first two reasons in the following.
Externalities are an important determinant of the costs and benefits of delegation of
discretion. With respect to the benefit side, externalities may enter to the extent that increased
delegation is accompanied by, for example, increased knowledge-sharing (Osterloh and Frey
2000; Mudambi et al. 2003). With respect to the cost side, externalities enter, for example, in
the form of the moral hazard, duplicative efforts, and misuse of corporate resources that may
result from delegation. Thus, although delegation may be a part of an attempt to make the
organization more modular (Zenger 2002), complex interdependencies are not necessarily
eliminated by delegation per se. Indeed, delegation of discretion may sometimes introduce

7
more interdependencies between organizational units (e.g., when delegation implies more
horizontal links between units). Also, delegation may imply more extensive rights to draw on
corporate resources (as in a matrix organization).
Because complex interdependencies still exist under delegated discretion, major
changes in contingencies are likely to change the optimum degree of delegation. For
example, changes in the firm’s overall strategy may require the building-up of a new product
platform. Such new technologies typically require the delegation of more discretion to
designers and engineers in order to stimulate exploration through wide bandwidth
communication channels. Or, a change in the competitive conditions, such as an impending
price war, may dictate that discretion be diminished in order to curb slack and reduce costs.
Many contingencies cannot be foreseen, or it is too costly to try to do so (Malmgren 1961;
Williamson 1996). Moreover, how exactly contingencies impact on the preferred level of
delegation may also be difficult to specify ex ante (Coase 1937). This introduces a need for ex
post decision-making (Coase 1937; Malmgren 1961; Williamson 1996). Centralized
decision-making, that is, discretionary authority, becomes a preferable mechanism of
coordination when those who may hold authority has a superior understanding of how
contingencies influence interdependencies and how this impacts on the preferred degree of
delegation.
The preferred delegation of discretion may also change because of changed managerial
perceptions, even if no outside contingencies change. For example, a change in the
management team may cause the team’s “image” (Penrose 1959) of the firm’s opportunity set
to change. Or, managers may develop certain cognitive biases that twist their assessment of
costs and benefits (Bazerman 1994), leading them to change their perception of what the
optimal trade-off is. For example, under what psychologists call the “loss aversion bias,” a
loss relative to the status quo is seen as more undesirable than a gain relative to the same
status quo is seen as desirable (e.g., Kahneman et al. 1991). This may lead managers to
overestimate costs relative to benefits, which will cause them to change the degree of
delegation of discretion.
These psychological effects may be aggravated by overconfidence biases. Robust
findings in experimental psychology show the presence of a systematic overconfidence bias in
judgment, that is, people tend to trust their own judgments more than is “objectively”
warranted. Managers are not likely to be exceptions to this bias, perhaps quite the contrary.

8
The presence of the overconfidence bias in the judgments that underlie managerial decision-
making may strengthen managers’ incentive to change the level of delegation of discretion.
Changing delegation of discretion may not, as argued earlier, in itself be harmful to employee
motivation, namely when employees and management basically agree on the need for a
change. However, the presence of biases in the perceptions and judgments of the parties
means that it is harder for employees to ascertain whether intervention takes place for bad or
for good causes, that is, whether or not there is a break with established implicit contracts and
commitments to refrain from opportunistic intervention. We consider opportunistic
managerial intervention in the following section.
Opportunistic Managerial Intervention and Employee Motivation
Williamson’s (1996) distinction between intervention for good cause and intervention for
bad cause (i.e., opportunistic managerial intervention) is a fundamental one, because it directs
attention to the benefits as well as the costs of managerial authority. In terms of Williamson’s
distinction the preceding examples of managerial intervention largely fall in the category of
intervention for good cause, although we have introduced a perceptual and cognitive issue,
namely, employees may mistake good for bad causes and vice versa, that is not present in
Williamson’s discussion. More generally, employee motivation is arguably mediated by
employee perceptions of what motivates managerial intervention and whether and in which
manner managerial intervention breaks with existing psychological contracts (Rousseau 1989;
Robinson and Morrison 1995; Coyle-Shapiro and Kessler 2000). Thus, intervention that
essentially harms employees (e.g., leads to layoffs) may still not harm motivation, particularly in
a time of a severe organizational crisis and given that management succeeds in convincing
employees of the need for layoffs. Or, managerial intervention that objectively benefits all
relevant parties may be perceived by employees as breaking with psychological contracts. In
sum, in ascertaining the nature of managerial intervention, employees face a complicated signal
extraction problem. Our focus is on managerial intervention that is perceived by employees as
“intervention for bad cause,” that is, “opportunistic managerial intervention.”4
The relevant organizational behavior literature, which is largely based on social
psychology (Argyris 1960; Rousseau 1989; Robinson and Morrison 1995; Coyle-Shapiro and
4 Note that this is not entirely congruent with the notion of opportunism in Williamson (1996), primarily because
Williamson does not incorporate the perceptual issues that we do, and therefore does not allow for difficulties of
distinguishing between what is and what is not opportunistic behavior.

9
Kessler 2000), suggests that such managerial intervention amount to, in economic terms,
reneging on implicit contracts or explicit commitments. For example, managers may overrule
employee decisions that are made on the basis of delegated decision rights, or managers may
renege on the level of delegation itself. As further suggested by the relevant organizational
behavior literature (e.g., Robinson and Rousseau 1994; Rousseau 1989), loss of motivation
results. In particular, organizational citizenship behavior  that is, employee behavior that
promotes organizational efficiency but is not (perhaps, cannot be) explicitly recognized by an
organization’s reward system  may suffer from opportunistic managerial intervention
(Robinson and Morrison 1995).
The psychological literature on cognitive biases suggests further reasons why
motivation may be harmed by opportunistic managerial intervention. In an employee
relationship, employees develop implicit and explicit expectations of the contract governing
the relationship (Coyle-Shapiro and Kessler 2000), and particularly of the benefits that they
believe they deserve under the implicit contract, that is, their “entitlements” (Heath et al.
1993). For example, certain levels of delegated discretion may become “status quo” points, in
the sense that they represent what employees believe are their entitlements. Thus, if employs
enjoy considerable discretion this may become part of their (perceived) entitlements. As
discussed earlier, loss aversion implies that a loss relative to the status quo point is seen as
more undesirable than a gain relative to the same point is seen as desirable. This means that
employees will develop a bias against changing the level of discretion in a downwards
direction, and that they can be expected to resist such changes, as well as suffer a loss of
motivation if the change is, in fact, forced upon them. The above reasoning is summed up in
the following hypothesis:
Hypothesis 3: Employee motivation varies negatively with opportunistic managerial
intervention.
For the firm, this is a problem to the extent that loss of motivation leads to employees cutting
back on the effort they supply to the firm, and also on their firm-specific investments in
human capital.5 This implies the following hypothesis:
5 Of course, loss of motivation may not automatically lead to, for example, less effort supply, if monitoring
systems or extrinsic motivation can substitute for the loss of motivation caused by opportunistic managerial
intervention.

10
Hypothesis 4: Overall firm performance varies negatively with opportunistic managerial
intervention.
Why Opportunistic Managerial Intervention?
As suggested earlier, the problem of loss of motivation because of opportunistic managerial
intervention is related to what Oliver Williamson (1996: 150) calls the “impossibility of selective
intervention,” that is, the puzzle of “Why can’t a large firm do everything that a collection of
small firms can and more?” Thus, a large firm could replicate the market and only selectively
intervene when there would be expected net gains from this, so that “… the firm will do at least
as well as, and will sometimes do better than, the market.” However, Williamson points out
argues that such selective intervention is ”impossible.” Incentives are diluted, because the option
to intervene ”… can be exercised both for good cause (to support expected net gains) and for bad
(to support the subgoals of the intervenor)” (Williamson 1996: 150-151), and employees know
this. Promises to only intervene for good cause can never be credible, Williamson argues,
because they are not enforceable in a court of law. A fundamental problem  in theory as well
as managerial practice  is therefore how to maximize managerial intervention for “good cause,”
while avoiding intervention for “bad cause.” Our discussion provides a further reason why first-
best selective intervention is impossible: Employees may have difficulties distinguishing between
intervention for good and bad cause.
However, apart from these perceptual problems it is not immediately apparent why
opportunistic managerial intervention should ever take place. According to Hypothesis 4
opportunistic managerial intervention destroys value. However, there are least two explanations
for why value-destroying opportunistic managerial intervention may take place, namely
managerial private benefits and managerial time inconsistency.
According to the first explanation, managers may derive a private benefit (in whatever
form) from managerial intervention that destroys value, when organizational and private costs
and benefits are timed in certain ways. For example, managers who are up for promotion may
derive private benefits from imposing restrictions to strongly cut the costs of the slack and
spillover effects associated with a high level of delegation of discretion. If organizational
benefits follow later than these costs, managers have an incentive to engage in managerial
intervention that harms motivation. The organizational costs of such actions may not be borne

11
by the managers themselves, for example, because they may have left the firm or the position in
favor of another firm or position.6
The explanation from managerial time inconsistency relates to a familiar problem in
political economy (Weingast and Marshall 1988; Moe 1997). Typically, this problem starts out
from a timing of costs and benefits that is the opposite of the one in the above explanation. For
example, governments have an incentive to initially promise not to confiscate (too much of) the
wealth created by entrepreneurs in order to strengthen their incentives to actually undertake
investments, and then, in some later period, deviate from this promise and confiscate substantial
portions of the created wealth. In the context of delegation, this kind of behavior may consist in,
first, promising substantial discretion. When employees, enthused about their new extended
discretion, come up with profit-improving ideas about how to improve products, processes, etc.,
managers may harvest these, decide that the organization already has its hands full with
implementing the ideas, and that the level of delegated discretion may be usefully reduced in
order to save costs.7 However, the political economy literature referred to above also suggests
that these problems may be checked by various institutions and mechanisms. We consider these
next.
Credible Delegation
The political economy concept of credible commitment (see also Williamson 1996)
implies that it is often in an organization’s long-term interest to avoid later period actions that
break promises (with respect to delegation), thereby harming organizational members, and that
avoiding such behavior may be accomplished by credibly constraining the flexibility of
managers in such a manner that the initial promise becomes credible (Weingast and Marshall
1988; Moe 1997). In the present context, there are two classes of ways in which promises to not
engage in opportunistic managerial intervention may be made credible, namely what may be
called internal and external mechanisms.
With respect to internal mechanisms, managers may stake their personal reputations
(Miller 1992; Argyres and Mui 1999), for example, through symbolic and communicative acts,
6 Even if managers are in fact made partly responsible for later organizational costs, their rate of time preference may
be such that these costs are heavily discounted.
7 This may help explain why organizations often “vacillate” between loose and hierarchical structures
(Nickerson and Zenger 2000).

12
for example, announcing in large-scale company gatherings one’s firm commitment to certain
policies and values (Brockner et al. 1992). This suggests the following hypothesis:
Hypothesis 5a: Opportunistic managerial intervention varies negatively with the strength
of managers’ personal reputations for pursuing a “fair” or “hands off” policy in dealing
with employees.
It is well know that, in general, reputation effects are far from perfect with respect to
constraining opportunistic behaviors (Williamson 1996). This also holds for reputation effects
inside the hierarchy. For example, managers change jobs and may not carry their reputation with
them. Corporate cultures are longer lasting than personal reputations and serve to enforce
implicit contracts in situations where personal reputations fail (Kreps 1990):
Hypothesis 5b: Opportunistic managerial intervention varies negatively with the extent to
which corporate culture implies expectations that managers will pursue a “fair” or
“hands off” policy in dealing with employees.
Hierarchical structure also plays a role in constraining managerial opportunistic
intervention. Thus, Milgrom (1988) argues that employee rent-seeking that aims at influencing
hierarchical superiors to selective intervene to the benefit of the rent-seeking employees will be
constrained by rigid, hierarchical structures which makes such rent-seeking more costly. Also,
upper and lower-level managers may differ in their preferences for intervention, for example,
lower-level managers may derive a private benefit from overruling, whereas upper-level
managers do not (Aghion and Tirole 1997).
A third reason why hierarchical structure may constrain opportunistic managerial
intervention (in fact, all managerial intervention) is that the hierarchy is not just a structure of
authority, but also one of information (Thompson 1967; Galbraith 1974). Thus, there will be an
informational distance between those possessing authority and those to whom discretion has
been delegated. The size of this informational distance influences the basis for exercising
judgment with respect to decisions whether to overrule employees or not. All else being equal,
the more hierarchical layers that information has to pass through before reaching the level
exercising authority, the less adequate is this basis likely to be. Moreover, even though there
may be few hierarchical layers, managerial task descriptions may be such that managers will
essentially be overloaded if they insist on being sufficiently informed to be in a position to
overrule. If the manager realizes that because of information overload, he is not in a position to

13
rationally decide whether to overrule or not, he should not overrule (Aghion and Tirole 1997).
Thus, this reasoning predicts that overruling of employees is less likely to occur in organizations
with large informational distances and/or managers that are heavily burdened with information:
Hypothesis 5c: Opportunistic managerial intervention varies negatively with the
informational distance in the corporate hierarchy.
Some employees or groups of employees may be particularly costly for management to
overrule, because they control critical resources, notably their own human capital. For example,
Henry Ford II and the rest of the Ford top management team tolerated the open disagreement
with official Ford strategy expressed by Lee Iacocca and his clique of loyal managers, because of
the marketing skills exercised by Iacocca and his men (Halberstam 1986). Overruling such
employees means that they may cut back on the supply of their essential services and may refrain
from augmenting their valuable human capital. This suggests the following hypothesis:
Hypothesis 5d: Opportunistic managerial intervention varies negatively with the degree
of human capital specificity.
Employees with strongly specialized, important human capital may possess considerable
bargaining power and influence (Rajan and Zingales 1998). However, such influence may also
be secured through other means, such as extensive employee ownership of the firm. This means
that employee interests may be more strongly reflected in corporate decision-making, implying
that in such firms, opportunistic managerial intervention may be less prevalent:
Hypothesis 5e: Opportunistic managerial intervention varies negatively with the degree to
which employee interests are represented in corporate decision-making.
With respect to external mechanisms that may enforce delegated discretion, a clear
example is strong trade unions or professional associations. Their influence may imply that
certain rights are so strongly protected (i.e., they are outside the “zone of acceptance,” Simon
1951) that management cannot realistically change these (Argyres and Liebeskind 1999).
Hypothesis 5f: Opportunistic managerial intervention varies negatively with the degree of
unionization and the strength of unions and professional associations.
In the following section, we present our data set, the methods we have applied, and the
results.

14
III. Data, Variables, Constructs, and Results
Data Collection
Data were collected by mail questionnaire after an initial pilot testing of the instrument.
The sample population is composed of all firms in the Spanish food and electric/electronic
industries (SIC 20 and SIC 36) with a turnover of 3 million euros or more in the year 20008.
Following these criteria, the population of the study was drawn from the directory, DB
Marketing: 700.000 empresas españolas. This directory is updated on an annual basis by the
international management consultancy, Dun & Bradstreet. From this database we identified
3.040 firms that met the conditions described above. We mailed an initial questionnaire with a
customized letter addressed to the production manager in each firm. 36 questionnaires were
returned, because either the address was wrong or the firm had quitted its activity. Furthermore,
not all the remaining questionnaires were valid: Missing values and the unfeasibility of
identifying the firm to which some of the questionnaires belonged resulted in the final sample
being composed of 329 firms (11% of the total population). Assuming the worst scenario for a
binary variable, where [p = q = 50%], and imposing a confidence level of 95%, these figures
represent a sampling error of ±5’09%.
Variables and Constructs
Table 1 shows a brief description of the variables. Some of them required direct figures
from key respondents, while others have been addressed through the linear combination (using
Principal Component Analysis) of several indicators generally valued on a five point Likert-type
scale.
XXXXXXXX INSERT TABLE 1 ABOUT HERE XXXXXXXX
The process of building these variables followed two steps. First of all, and based on the
extant literature (Mowday, Steers and Porter 1982; Dewar, Whetten and Boje 1980; Lawrence
and Lorsch 1967; Pugh and Hickson 1976; Dow 1987; Grimshaw and Rubery 1998), a list of
indicators for each variable was presented to a group of three production managers and two
8 The kind of information required for this study is not usually available for smaller firms or results are often
rather obvious. Moreover, the greater the size of the firm, the more experience firms have and the higher the
qualification of the participant regarding the concepts included in the questionnaire; this obviously affects the
reliability of the responses by making the answers more rigorous.

15
operators from diverse firms.9 They were asked to discuss how representative each indicator was
of the corresponding construct and propose others that were not in the original list, concluding
with a different number of indicators for each variable that, in their view, reasonably reflected
what the variable tried to grasp. In a second stage, these indicators were tested in twenty
interviews together with the rest of the items of the questionnaire. We finally chose those for
each variable that provided not only the highest Cronbach’s α, but also a first component
through Principal Component Analysis that could explain more than 50% of the variance of the
items.
This process turned out to be satisfactory, although obviously not perfect. Thus, for
corporate culture and managerial opportunism, one or more indicators had to be pulled out in
order to get a better reliability of the scale (Cronbach’s α should be above 0,7 for a non-
exploratory analysis; cf. Nunnally 1978). In the case of motivation and delegation of discretion,
our initial measures based on the literature (Mowday, Steers and Porter 1982, and Dewar,
Whetten and Boje 1980, respectively) did not offer a first component that could explain more
than 50% of the variance of the items used in the preliminary test of the survey, so we decided to
stick to only one item. Thus, once the indicators were chosen for each construct and data was
available, Principal Component Analysis was applied for corporate culture and managerial
opportunism in order to get a single value. Table 2 shows the main figures.
XXXXXXXX INSERT TABLE 2 ABOUT HERE XXXXXXXX
Regarding the validity of the constructs, it is worth noting two issues. First, concerning
their uni-dimensionality, we can see in Table 2 that the component that has been extracted can
explain most of the variance of the items. Second, searching for the most sensible way to gather
a representative collection of items for each latent variable, we resorted to a wide range of
seminal contributions on the measurement of organizational traits as a first step to build the
scales. The items used for corporate culture were initially extracted from Lawrence and Lorsch
(1967), Pugh and Hickson (1976) and Kotter and Heskett’s (1992). Although we did not use an
explicit measurement scale on which to mould the construct of “managerial opportunism,” the
initial items have been definitely inspired by the insights and specific examples contained in such
works as Willman (1983), Dow (1987), Rousseau (1989), Miller (1992), and Grimshaw and
9 We considered this step crucial, even for those measures that have previously been tested. This is because the
translation of scales from English to Spanish may change the perception of the respondent. Moreover, the items
themselves may not make sense for cultural reasons.

16
Rubery (1998). Finally, and for strictly operational reasons (particularly to search for an easier
interpretation of absolute figures), the latent variables were subsequently transformed to make
them start with 1. The algorithm is:
1) yi* = –(minimum value of y) + 1 + yi
Finally, some of the variables referred to in the above hypotheses were not directly
measured. This is the case for the “strength of managers’ personal reputations” construct. We
proxy this construct with the age of the firm variable, based on the argument that young firms
have higher expected mortality, which implies that the value of a manager’s reputation in
such a firm is smaller than in an older firm with a lower expected mortality. Also, we did not
directly measure the construct “informational distance in the corporate hierarchy” (H5c). We
proxy this construct with the size variable, because it is reasonable to expect a positive
relation between the size of a firm and the depth of its corporate hierarchy. Finally, the
variable “degree to which employee interests are represented in corporate decision-making”
(H5e) has been measured by the hierarchical form of the firm, which distinguishes between
capitalist and worker-owned firms.
Empirical Results and Discussion
Table 3 shows the means, standard deviations and correlations for the variables. Two
initial important insights have to do with the rather low level of delegation of discretion that
we can find in our study, whereas managerial opportunism achieves a high figure in average
terms. Moreover, both are negatively correlated. Regarding their association with other
variables, a high delegation of discretion is strongly related to a strong corporate culture,
workforce motivation and labor productivity. On the other hand, managerial opportunism
appears to be associated with little human specificity involved in the labor transaction, lower
motivation and socialization, and is especially relevant in SMEs and capitalist firms when
compared to large corporations and worker owned enterprises, respectively.
XXXXXXXX INSERT TABLE 3 ABOUT HERE XXXXXXXX
Since the object of our investigation has to do with the interaction among managerial
opportunism, motivation, and performance, we have developed a path analysis which,
compared to conventional multivariate techniques, allows us to design a model with various

17
levels of dependency. Although probably far from being exhaustive in terms of including all
potentially relevant independent variables, the model does seem to be the best one which our
insights allowed us to construct prior to this research. Thus, our aim is not only to verify or
refute each one of the above hypotheses separately, but also to test whether their interaction is
statistically significant10.
Path analysis assumes that relations among variables are linear, residuals from the
regressions are not correlated among them and variables are measured without error (Bagozzi
1982; Bollen and Long 1993).11 Given these assumptions and for the sake of simplicity, we
can start by designing a diagram that reflects the relations of dependence among the variables
that are included in our hypotheses. Next, we will convert this diagram into a system of
simultaneous equations: this constitutes our structural model with the several path coefficients
that we estimate here. Third, we evaluate the model so that the possibility to re-specify it and
thus achieve a better goodness of fit is assessed.12 Finally, we interpret and evaluate the final
model.
Following this scheme, Figure 1 presents the path diagram with the relations that our
propositions suggest. First, observe that the significant correlations among the exogenous
variables shown in Table 3 have been represented in the diagram by the two-headed arrows.
Note that these correlations are estimates of the population correlation matrix of the
independent variables in the model; this is the reason why there are some slight differences
between these figures presented in Figure 1 and the ones presented in Table 3. Anyhow, since
no multicolinearity problems were identified, all correlations among the exogenous variables
are maintained.
Concerning the endogenous variables and starting with opportunistic managerial
10 Construct building (managerial opportunism and corporate culture) could also be implemented in the same
model. We chose to do it separately because 1) the results do not change and 2) the final model is thus made
considerably simpler.
11 On why motivation is treated as a continuous dependent variable see, for instance, Bohrnestedt and Borgatta
(1981). Their argument, which we share, is that the consequences of assuming that data are interval when in fact
they are ordinal are so small in most cases that the gain in statistical elegance and power justifies the possible
distortion.
12 There are dozens of measures of the goodness of fit (GF) for this kind of models, and they are generally
grouped under three headings: Absolute GF, Incremental GF and Parsimonious GF (Bentler and Bonnet 1980;
Bollen 1989; Bollen and Long 1993; Bagozzi 1982, 1991). The measures used here are the ones which appear
to be more widespread in the empirical literature and in the specific software packages design for this tool (EQS,
LISREL, AMOS, etc.). Moreover, following Hair et al. (1998), these measures have been chosen ex ante, i.e.,
before performing the estimation.

18
intervention, observe that, as suggested by hypothesis 5, it varies negatively with the level of
expert knowledge involved in the labor transactions (H5d), the strength of corporate culture
(H5b), the degree of unionization (H5f), and the size (H5c), age (H5a) and hierarchical form
of the firm (H5e).
Next, consider motivation, and observe that the one-headed arrows represent its
dependency on the level of delegation of discretion (H1) and the degree of managerial
opportunism (H3). The latter here also acts as an independent variable, just as it happens also
in the case of delegation of discretion, which is supposed to decline as managerial
opportunism increases.
Additionally, hypotheses 2 and 4 propose that, despite the fact that firm performance
obviously depends on many other variables, it will be affected by both the level of delegation
of discretion and managerial opportunism.
Note, finally, that there are also two second order relations involving motivation and
performance, on the one hand, and managerial opportunism and delegation of discretion, on
the other. They do not explicitly appear in our set of hypotheses, mainly because these
interactions have already been well established in previous literature; in the first case, for
instance, in the such studies as McClelland (1955), Herzberg et al. (1959), and Vroom (1964),
and in the more recent ones by Prokopenko (1987) or Frey and Osterloh (2002). Regarding
the relation of managerial opportunism with delegation of discretion, the rationale is that, no
matter how short termed decisions might be (for instance working overtime, changing shifts,
assuming new tasks, etc.), delegating discretion restricts the ability of managers to go beyond
the ex ante agreed “zone of acceptance” in the sense described by Simon (1951), Willman
(1983), Dow (1987) and Kreps (1990). Therefore, delegation cannot be credibly sustained in
firms where managerial opportunism is high. Thus, the initial structural model takes the
following form:
(1) man_opp = α1+ β14 hum_spec + β17 socializ + β18 size + β16 age + β15 prop_uni + β13 hf + e1
(2) perform = α2+ β22 man_opp + β29 motiv + β21 del_disc + e2
(3) motiv = α3+ β32 man_opp + β31 del_disc + e3
(4) del_disc = α4+ β42 man_opp + e4
XXXXXXXX INSERT FIGURE 1 ABOUT HERE XXXXXXXX

19
The path coefficients of the former model are the main object of our estimation; they
represent the beta weights obtained from a set of multiple regressions on the posited
relationships within the model. In this case, given the absence of multivariate normality and
the size of the sample, the method of estimation has been based on the Maximum Likelihood
criterion with a bootstrap of 200 sub-samples.13
Results are shown in Table 4. The path coefficient reflecting the influence of corporate
culture on managerial opportunism does not seem to be significant. The overall measures for
the goodness of fit, on the other hand, reveal rather ambiguous values. Thus, although the
GFI achieves a satisfactory figure (above 0.9), the rest of the measures are rather low (AGFI,
TLI, NFI, PNFI and PGFI) or offer unacceptable values (Chi-square probability, RMSEA and
AIC).
XXXXXXXX INSERT TABLE 4 ABOUT HERE XXXXXXXX
These considerations call for a reformulation of the model in order to achieve a better
goodness of fit and check the possible influence that including non-significant variables in the
model might exert on the rest of the path coefficients, which could eventually become non-
significant or suffer important alterations. Hence, Figure 2 shows a new diagram in which
corporate culture has been taken out, while the rest of the relationships have been retained in
the way showed by our re-specified structural model:
(1) man_opp = α1+ β14 hum_spec + β18 size + β16 age + β15 prop_uni + β13 hf + e1
(2) perform = α2+ β22 man_opp + β29 motiv + β21 del_disc + e2
(3) motiv = α3+ β32 man_opp + β31 del_disc + e3
(4) del_disc = α4+ β42 man_opp + e4
XXXXXXXX INSERT FIGURE 2 ABOUT HERE XXXXXXXX
Table 5 shows the new results. In effect, all of the estimators seem now to be
significant and the fit of the model achieves more than acceptable values except for the Chi-
13 A maximum likelihood estimation alone would require multivariate normality. In order to solve this problem
(given that our data exhibits a high kurtosis), bootstrapping extracts several random sub-samples and calculates
the mean of the estimators for each one of them. Other estimation methods –also valid- like generalized or
unweighted least squares are less demanding in terms of parametric assumptions, while taking out the cases
affecting the kurtosis would harm our sample representativeness.

20
square test (reflecting whether there exist significant differences between the observed and the
reproduced covariance matrix). In this case, Bollen (1989) has nevertheless shown that the
higher the size of the sample, the worse the goodness of fit (an ideal size would be between
100 and 200 cases). Since our sample contains 329 units, this might be the reason why the
probability of the Chi-square is not significant; note that the rest of the measures, nonetheless,
confirm that the model is significant.
With respect to the interpretation of the model, it appears that when compared to
smaller, younger and capitalist firms, the level of opportunistic managerial intervention
becomes lower in large corporations, older firms and cooperatives, respectively. Additonally,
the propensity of managers to behave opportunistically seems also higher in firms with low
specific human assets and with a low level of union affiliation. These findings confirm
hypotheses 5a, 5c, 5d, 5e and 5f, which state a negative relation between opportunistic
managerial intervention and, respectively, managers’ personal reputations (proxied by age),
the informational distance in the corporate hierarchy (proxied by size), the level of human
capital specificity, the degree to which employee interests are represented in corporate
decision-making (proxied by hierarchical form), and finally, the strength of unions and
professional associations (proxied by union affiliation).
Moreover, the standardized coefficients convey information for assessing the relative
influence that each one of the independent variables exerts on managerial opportunism. Thus,
the hierarchical form of the firm seems to be the main mechanism that helps to avoid
opportunistic intervention on the part of managers; that is, the higher the degree to which
employee interests are represented in corporate decision-making, the more difficult it is for
managers to implement oportunistic interventions. In fact, it is even more important than the
bargaining power stemming from local and specific knowledge or from the strength of unions
in the firm. The size or the age of the firm, in turn, although significant, apparently explain a
lower percentage of the variance of the dependent variable.
XXXXXXXX INSERT TABLE 5 ABOUT HERE XXXXXXXX
Regarding motivation, on the other hand, the results show that (1) workers seem to be
more motivated in firms with a higher delegation of discretion; (2) as the level of managerial
opportunism increases workforce motivation clearly goes down; and (3) there is a significant
indirect effect of managerial opportunistic intervention on motivation through delegation of

21
discretion. This evidence suggests that we cannot refute hypotheses 1 and 3 linking employee
motivation to delegation of discretion and managerial opportunistic intervention.
With respect to firm performance, observe that Figure 2 and Table 5 reflect also two
additional important facts: to begin with, they confirm Hypotheses 2 and 4 which suggest a
direct influence of delegation of discretion and managerial opportunistic intervention on firm
performance. And secondly, they verify the relevant indirect effects that, regarding workers’
productivity, both managerial opportunism and delegation of discretion pose: thus, while the
latter exerts an indirect effect through motivation (Hypothesis 1), the former does it not only
through motivation (Hypothesis 3) but also through delegation of discretion itself (second
order relation).
V. Concluding Discussion
In this final section, we sum up how we have contributed to existing theory, discuss
limitations of the study and suggest implications for future research.
Contribution to Established Literature
Most firms make use of both authority and delegated discretion. However, the main
point in this paper is that this gives rise to a latent conflict. The problem arises because
“contracts” to delegate discretion are not enforceable in a court of law. Credible delegation
may therefore be hard to sustain. However, we have pointed to and analyzed how various
mechanisms may make delegation credible.
Although this set of issues are not neglected in the theory of the firm literature per se
(e.g. Miller 1992; Aghion and Tirole 1997; Baker, Gibbons and Murphy 1999; Falaschetti
2002), it is still fair to say that they have been given relatively little attention in this body of
work. One manifestation of this is that economics of organization analyses of “opportunism”
has rather exclusively dealt with employee opportunism (Williamson 1996), employer
opportunism being almost entirely neglected (cf. Dow 1987). To be sure, the basic idea that
we have elaborated in this paper may be argued to be present already in Milgrom’s (1988)
argument that organizational form partly reflects an attempt to cope with employee rent-
seeking and the inefficient selective intervention that may result from such rent-seeking.
Also, a number of recent organizational economics contributions clearly go quite some way
towards understanding the incentive liabilities of centralized authority (e.g., Milgrom and

22
Roberts 1996; Aghion and Tirole 1997; Baker, Gibbons and Murphy 1999). However, this
remains an under-researched area in the economics of organization literature, particularly
given the apparently high incidence of managerial opportunism (cf. Coyle-Shapiro and
Kessler 2000).
In contrast to the organizational economics literature, much of the organizational
behavior literature on psychological contracts, organizational citizenship behavior and the like
is very strongly empirical. However, this literature does not explicitly frame the issues in
rational choice terms. Still, this literature is considerably more detailed with respect to
analyzing the actual contents of psychological (implicit) contracts between those who hold
discretionary authority and those who do not and the psychological mechanisms that are at
work in the case of perceived contract breach. We have mainly used this literature to provide
support for some parts of our hypothesis development. However, we conjecture that the
organizational behavior literature in this field and the relevant organizational economics
literature may well enter a fruitful liaison.
Limitations
A number of inherent limitations of the dataset imply that our analysis is far from
perfect. First, as in most studies, some of our proxies reflect a certain roughness derived from
data availability and reliability. Since their validity has been justified on theoretical as well as
on empirical grounds, nevertheless, we think that they reasonably represent and capture the
theoretical constructs they proxy.
Second, the limitations of the data set have constrained our theoretical framework. For
example, we argue that employee motivation positively depends on the degree of delegation.
The link between delegation (or “task autonomy”) and motivation has long been recognized
in social psychology (e.g., Roethlisberger et al. 1939). It has also long been recognized that
for delegation to be effective, employees need to have not only the opportunity but also the
incentive and the ability to engage in self-management. We have argued that reductions in
delegation, at least when these are perceived as reflecting managerial opportunism, lead
employees to reduce effort and human capital investment. This may lead to a confusion of
cause and effect. For example, it is conceivable that cutting back on delegation is a result of
finding out that employees lack the skills that are necessary to engage in self-management.

23
Thirdly, more generally, much of our reasoning admittedly proceeds in dynamic terms
— for example, we make references to breaking psychological contracts — that do not
correspond directly to the measures that we use (e.g., we don’t measure the incidence of
broken contracts) and the cross-sectional nature of the study.
Implications for Future Research
Future research may well start from some of the above limitations. Thus, panel data
need to be collected so as to better correspond to the dynamic nature of the argument. Also,
it would be desirable if data allowed for cross-country comparisons. Otherwise, we cannot
rule out the possibility of a country bias in our results.
Our study also suggests a number of avenues for further theoretical research. An
obvious route is to formalize our verbal argument. More substantively, there are theoretical
implications that await further development. One such implication is that the problem of
reducing opportunistic managerial intervention may differ systematically across firms,
depending on the details of their internal structure so that some organizational forms are
systematically more heavily burdened with problems of opportunistic managerial
intervention. Another implication is that the discussion in this paper relates to the classic
issue of the determinants of the boundaries of the firm. Thus, a fundamental premise of the
analysis in this paper is that in firms delegated decision rights are loaned, not owned (Baker,
Gibbons, and Murphy 1999). Ultimate decision-making rights can only be transferred from
bosses to subordinates by transferring ownership (i.e., creating a new firm). The problem of
sustaining credible delegation stems from this basic difference in ownership. The analysis in
this paper thus makes direct contact with those modern theories of economic organization
(Hart 1995; Williamson 1996) that stresses the importance of ownership for understanding the
boundaries of the firm. Finally, we have pointed to the desirability of more fully integrating
organizational behavior perspectives on psychological contracts with organizational
economics ideas, in order to get a fuller and more relevant understanding of the workings and
implications of psychological and implicit contracts. Both fields stand to benefit from such
an exercise (Gibbons 1999).

24
References
Aghion, Philippe and Jean Tirole. 1997. “Formal and Real Authority in Organization,” Journal
of Political Economy 105: 1-29.
Argyris, Chris. 1960. Understanding Organizational Behavior. Homewood, IL: Dorsey Press.
Argyris, Nicholas and Julia Porter Liebeskind. 1999. “Contractual Commitments, Bargaining
Power, and Governance Inseparability: Incorporating History into Transaction Cost
Theory,” Academy of Management Review 24: 49-63.
Argyris, Nicholas and Vai-Lam Mui. 1999. “A Political-Economic Approach to Organizational
Dissent,” Working Paper.
Armstrong, Mark. 1994. “Delegation and Discretion”. Discussion Paper in Economics and
Econometrics, Department of Economics, University of Southampton.
Bagozzi, R.P., ed. 1982. “Special Issue on Causal Modeling.” Journal of Marketing Research 19
(4).
Bagozzi, R.P. 1991. Structural Equation Models in Marketing Research, Chicago: American
Marketing Association.
Baker, George, Robert Gibbons, and Kevin J. Murphy. 1999. ”Informal Authority in
Organizations,” Journal of Law, Economics and Organization 15: 56-73.
Barzel, Yoram. 1997. Economic Analysis of Property Rights. 2nd ed. Cambridge: Cambridge
University Press.
Bazerman, Max H. 1994. Judgment in Managerial Decision Making. New York: Wiley.
Bentler, P.M. and D.G. Bonnet. 1980. “Significance tests and goodness of fit in the analysis of
covariance structures”, Psychological Bulleting 88: 588-606.
Bijl, Paul de. 1996. “Delegation of Responsibility in Organizations,” chapter 4 in idem. Essays in
Industrial Organization and Management Strategy. Phd thesis, Tilburg University.
Bollen, K. 1989. Structural equations with latent variables, New York: John Wiley and Sons.
Bollen, K. and Long, S. 1993. Testing Structural Equations Models, Thousand Oaks
(California): Sage.
Brockner, J., T.R. Tyler, Cooper, and R. Shneider. 1992, ”The Influence of Prior Commitment to
an Institution on Reactions to Perceived Infairness: The Higher They Are, the Harder
They Fall,” Administrative Science Quarterly 37: 241-61
Casson, Mark. 1994. “Why are Firms Hierarchical?,” International Journal of the Economics of
Business 1: 47-76.
Coase, Ronald H. 1937. “The Nature of the Firm,” in Nicolai J. Foss, ed. 1999. The Theory of the
Firm: Critical Perspectives in Business and Management, Vol II. London: Routledge.
Conger, J. and R. Kanungo. 1988. “The Empowerment Process: Integrating Theory and
Practice,” Academy of Management Review 13: 471-482.
Coyle-Shapiro, Jackie and Ian Kessler. 2000. “Consequences of the Psychological Contract for
the Employment Contract: A Large-Scale Survey,” Journal of Management Studies 37:
903-930.

25
Dewar, R., D. Whetten and D. Boje. 1980. “Examination of the Reliability and Validity of the
Aiken and Hage Scales of Centralization”, Administrative Science Quaterly 25: 120-128.
Dow, Gregory K. 1987. “The Function of Authority in Transaction Cost Economics,” Journal of
Economic Behavior and Organization 8: 13-38.
Falaschetti, Dino. 2002. “Golden Parachutes: Credible Commitments or Evidence of Shirking?,”
Journal of Corporate Finance 8: 159-178.
Fama, Eugene and Michael C. Jensen 1983. ”Separation of Ownership and Control,” Journal of
Law and Economics 26: 301-325.
Foss, Nicolai J. 1997. “On the Rationales of Corporate Headquarters”, Industrial and Corporate
Change 6: 313-339.
Frey, Bruno S. and Margit Osterloh. 2002. Succesful Management by Motivation: Balancing
Intrinsic and Extrinsic Incentives, Heidelberg and New York: Springer.
Galbraith, Jay R. 1974. “Organization Design: An Information Processing View,” Interfaces 4:
28-36.
Gal-Or, Esther and Raphael Amit. 1998. “Does Empowerment Lead to Higher Quality and
Profitability?,” Journal of Economic Behavior and Organization 36: 411-431.
Gibbons, Robert. 1999. ”Taking Coase Seriously,” Administrative Science Quarterly 44: 145-
157.
Grimshaw, D. and Jill Rubery. 1998. ”Integrating the Internal and External Labour Markets,”
Journal of Cambridge Economics 22: 199-220.
Hair, J. et al. .1998. Multivariate Data Analysis, Upper Saddle River (New Jersey): Prentice Hall
(5th edition).
Halberstam, David. 1986. The Reckoning. New York: Avon Books.
Harryson, Sigvald J. 2000. Managing Know-Who Based Companies. Cheltenham: Edward Elgar.
Hart, Oliver and Bengt Holmström. 2002. “Managerial Vision and Firm Scope,” mimeo.
Hart, Oliver. 1995. Firms, Contracts, and Financial Structure. Oxford: Oxford University Press.
Heath, Chip, Marc Knez, and Colin Camerer. 1993. “The Strategic Management of the
Entitlement Process in the Employment Relationship,” Strategic Management Journal
14: 75-93.
Hendry, John. 2002. “The Principal’s Other Problem: Honest Incompetence and the
Specification of Objectives,” Academy of Management Review 27: 98-113.
Herzberg, F. et al. 1959. The Motivation to Work. New York: John Wiley and Sons.
Holmström, Bengt and Paul Milgrom. 1994. “The Firm as an Incentive System,” American
Economic Review, 84, 972-991.
Holmström, Bengt. 1979. “Moral Hazard and Observability,” Bell Journal of Economics, 10,
74-91.
Holmström, Bengt. 1999. “The Firm as a Subeconomy,” Journal of Law, Economics, and
Organization 15: 74-102

26
Jensen, Michael C. and Karen Wruck. 1994. ”Science, Specific Knowledge and Total Quality
Management,” in Michael C. Jensen. 1998. Foundations of Organizational Strategy.
Cambridge: Harvard University Press.
Jensen, Michael C. and William H. Meckling. 1992. “Specific and General Knowledge and
Organizational Structure,” in Lars Werin og Hans Wijkander, eds. 1992. Contract
Economics. Oxford: Blackwell.
Kahneman, Daniel, Jack Knetsch, and Richard H. Thaler. 2001. “The Endowment Effect, Loss
Aversion and Status Quo Bias: Anomalies,” Journal of Economic Perspectives 5: 193-205.
Kotter, J.P., and J.L. Heskett. 1992. Corporate Culture and Performance, New York: Free
Press.
Kreps, David M. 1990. “Corporate Culture and Economic Theory,” in James G. Alt and Kenneth
Shepsle, eds. Perspectives on Positive Political Economy. Cambridge: Cambridge
University Press.
Lawrence, Paul and Jay Lorsch. 1967. Organization and Environment, Cambridge, MA: Harvard
Business School.
McMillan, G.S., R.D. Hamilton, and D. L. Deeds. 2000. “”Firm Management of Scientific
Information: An Empirical Update,” R&D Management 30: 177-182.
Malmgren, Harold B. 1961. “Information, Expectations, and the Theory of the Firm,” Quarterly
Journal of Economics 75: 399-421.
McClelland, D.C. 1955. Studies in Motivation. New York: Appleton-Century-Crofts.
Melumad, Nahum D., Dilip Mookherjee and Stefan Reichelstein. 1995. “Hierarchical
Decentralization of Incentive Contracts,” RAND Journal of Economcis 26: 654-672.
Miles, Raymond E., Charles C. Snow, John A. Mathews, Grant Miles and Henry J. Coleman, Jr.
1997. “Organizing in the Knowledge Age: Anticipating the Cellular Form,” Academy of
Management Executive 11: 7-20.
Milgrom, Paul. 1988. “Employment Contracts, Influence Activities, and Efficient
Organizational Design,” Journal of Political Economy 96: 42-60.
Milgrom, Paul and John Roberts. 1996. “Bargaining Costs, Influence Costs and the Organization
of Economic Activity,” in Louis Putterman and Randall Kroszner, eds. The Economic
Nature of the Firm. Cambridge: Cambridge University Press.
Miller, Gary. 1992. Managerial Dilemmas. Cambridge: Cambridge University Press.
Miller, Gary and Thomas Hammond. 1994. “Why Politics is More Fundamental Than
Economics,” Journal of Theoretical Politics 6: 5-26
Moe, Timothy. 1997. “A Positive Theory of Public Bureacracy,” in Dennis Mueller, ed. 1997.
Perspectives in Public Choice: A Handbook. New York: Cambridge University Press.
Mowday, R.T., R.M. Steers, and L.W. Porter. 1982. Employee-Organization Linkages:The
Psychology of Commitment, Absenteeism and Turnover, New York: Academic Press.
Mudambi, Ram, Susan M. Mudambi, and Pietro Navarra. 2003. “How to Motivate Knowledge
Workers: An Empirical Investigation of Motivation Crowding Theory,” Temple
University, The Fox School of Business and Management, Discussion Paper 716-03.
Nunnally, J.C. 1978. Psychometric Theory, New York: McGraw-Hill.

27
Osterloh, Margit and Bruno Frey. 2000. “Motivation, Knowledge Transfer and Organizational
Form,” Organization Science 11: 538–550.
Penrose, Edith T. 1959. The Theory of the Growth of the Firm. Oxford: Oxford University Press.
Prokopenko, J. 1987. Productivity Management. International Labor Organization.
Pugh, D.S. and D.J. Hickson. 1976. Organizational Structure in its Context: The Aston
Programme I, Farmborough: Saxon House/Lexington Books.
Purser, Ronald. 1998. The Self-Managing Organization. New York: The Free Press.
Rajan, Raghuram G, and Luigi Zingales 1998. “Power in a Theory of the Firm,” Quarterly
Journal of Economics : 387-432.
Robinson, Sandra L. and Denise M. Rousseau. 1994. ”Violating the Psychological Contract: Not
the Exception but the Norm,” Journal of Organizational Behavior 15: 245-259.
Robinson, Sandra L. and Elizabeth Wolfe Morrison. 1995. “Psychological Contracts and OCB:
The Effects of Unfulfilled Obligations on Civic Virtue Behavior,” Journal of
Organizational Behavior 16: 289-298.
Roethlisberger, F.J., Harold A. Wright, and William J. Dickson. 1939. Management and the
Worker. Cambridge, MA: Harvard University Press.
Root, Hilton. 1989. “Tying the King’s Hands: Royal Fiscal Policy During the Old Regime,”
Rationality and Society 1: 240-259.
Rousseau, Denise M. 1989. “Psychological and Implied Contracts in Organizations,” Employee
Responsibilities and Rights Journal 8: 121-139.
Sah, Raaj and Joseph E. Stiglitz. 1985. “Human Fallibility and Economic Organization,”
American Economic Review 75 (Papers and Proceedings): 292-297.
Simon, Herbert A. 1951. “A Formal Theory of the Employment Relationship,” in idem. 1982.
Models of Bounded Rationality. Cambridge: MIT Press.
Simon, Herbert A. 1991. “Organizations and Markets,” Journal of Economic Perspectives 5: 25-
44.
Tepper, Bennett J. and Edward C. Taylor. 2003. “Relationships among Supervisors and
Subordinates’ Procedural Justice Perceptions and Organizational Citizenship Behaviors,”
Academy of Management Journal 46: 97-105.
Thomas, K. and B. Velthouse. 1990. “Cognitive Elements of Empowerment: An Interpretive
Model and Intrinsic Task Motivation,” Academy of Management Review 15: 666-681.
Thompson, James D. 1967. Organizations in Action. New York: McGraw-Hill.
Vancil, Robert and Lee E. Buddrus. 1979. Decentralization: Managerial Ambiguity by Design.
Homewood, Il.: Irwin.
Vroom, V.H. 1964. Work and Motivation. New York: John Wiley and Sons.
Weingast, Barry and W. Marshall. 1988. “The Industrial Organization of Congress: Or, Why
Legislators, Like Firms, Are Not Organized as Markets,” Journal of Political Economy 96:
132-163.
Wernerfelt, Birger. 1997. On the Nature and Scope of the Firm: A Adjustment Cost Theory,”
Journal of Business 70: 489-514.

28
Williamson, Oliver E. 1970. Corporate Control and Business Behavior:
An Inquiry into the Effects of Organization form on Enterprise Behavior. Englewood
Cliffs: Prentice Hall.
Williamson, Oliver E. 1993. “Transaction Cost Economics Meets Posnerian Law and
Economics,” Journal of Institutional and Theoretical Economics 149: 99-118.
Williamson, Oliver E. 1996. The Mechanisms of Governance. Oxford: Oxford University
Press.
Willman, P. 1983. “The Organisational Failures Framework and Industrial Sociology”, in A.
Francis, J. Turk and P. Willman (eds.), Power, Efficiency and Institutions, London:
Heinemann.
Zenger, Todd. 2002. “Crafting Internal Hybrids,” International Journal of the Economics of
Business 9: 79-97.

TABLE 1
Description of the variables
Denomination and abbreviation Description
X1 Delegation of discretion (del_disc) Degree of delegation to adapt to eventual modifications on the production line that affect several work stations
(five levels).
X2 Managerial opportunism (man_opp) Propensity of a principal to impose orders exceeding the previously agreed limits to other agents who do not
have to be consulted. Construct built up with Principal Component Analysis (PCA).
X3
Hierarchical form (hf) Two values: 1 for capitalist firms and 2 for worker owned enterprises (most of the latter belong to the
Mondragon co-operative).
X4 Human specificity (hum_spec) Difference between the time that a new worker with no experience in the industry spends until she reaches the
normal productivity of her mates, and the time that a new worker who does have experience in the industry
spends until she reaches that normal productivity (five levels).
X5 Propensity to unionize (prop_uni)
Percentage of unionized workers within the firm.
X6Age (age) Three values: 1 for firms that have been in existence for 10 years or less; 2 for firms between 11 and 30 years,
and 3 for firms with more than 31 years.
X7Corporate culture (socializ) Extent to which certain norms and values are widely shared and intensely held throughout the organization.
Construct built up with PCA.
X8Size (size) Two values: 1 for SMEs with less than 100 operators; 2 for the rest.
X9Motivation (motiv) Degree to which workers’ commitment encourages them to do their best (five point Likert-type scale).
X10 Performance (perform) Value added divided by number of operators

30
TABLE 2
Summary of items retained in each construct built up with Principal Component Analysis
Var Items in the variable Factor
loading
KMO
index
Variance
explained
Cronbach's α
Managerial opportunism
Please indicate to what extent you agree with the following statements
(1 being ”strongly disagree” and 5 ”strongly agree”):
1.-If we paid overtime strictly, the firm would not be sustainable
2.-Some operators cannot always use up their holidays because of
production needs
3.-Flexibility and cost-saving requirements foster the use of short-term
contracts even for long term employment relationships
4.-Operators believe that managers press them excessively
,849
,782
,684
,704
,72 57,37 % ,74
Corporate culture
Please, indicate to what extent you agree with the following statements
(1 being ”strongly disagree” and 5 ”strongly agree”):
1.-Our operators know the history of the firm and its most important
achievements
2.-Our workers are acquainted with the firm’s short- and long-term
objectives
3.-Working in our firm makes our workers experience a sense of pride
,883
,884
,889
,72 76,06 % ,84

31
TABLE 3
Descriptive statistics and correlations
Min. Max. Mean St. dev. 1 2 3 4 5 6 7 8 9 10
1 del_disc 1 5 1,81 1,15 1
2 man_opp 1 5,35 3,46 1 -,24*** 1
3 hf 1 2 -- -- ,19*** -,44*** 1
4 hum_spec 1 5 2,48 ,972 ,16*** -,36*** ,020 1
5 prop_uni 0 100 31,89 26,44 ,075 -,127** -,42*** ,076 1
6 age 1 3 2,06 ,69 ,048 -,22*** -,1* ,062 ,43*** 1
7 socializ 1 5,54 3,56 1 ,51*** -,26*** ,38*** ,26*** -,139** -,088 1
8 size 1 2 -- -- ,081 -,26*** -,042 ,029 ,23*** ,31*** -,001 1
9 motiv 1 4 2,35 ,925 ,37*** -,41*** ,37*** ,24*** -,11** -,017 ,73*** -,02 1
10 perform -10,6 98,9 11,2 10,45 ,26*** -,33*** ,31*** ,264*** -,09* -,034 ,42*** ,071 ,51*** 1
a Pearson correlations for pairs of continuous variables and Spearman correlations when one or the two of them are ordinal
or categorical.
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level

32
Figure 1
Model 1 diagram with standardized estimates
age
socializ sizehum_spec prop_uni hf
del_disc
man_opp
motiv
perform
-,04
-,16
-,34
-,12 -,52
,37
,44
error
man_opp
error
perform
error
motiv
,31
-,31
-,20
,23
-,11 ,28 -,05
-,35
-,24
,30
error
del_disc
,30
,21
-,16

33
Figure 2
Model 2 diagram with standardized estimates
agesizehum_spec prop_uni hf
del_disc
man_opp
motiv
perform
-,16
-,34
-,11
-,53
-,16
,45
error
man_opp
error
perform
error
motiv
,30
-,32
-,20
,28 -,09
-,36
-,24 error
del_disc
,30
,30
,21

34
TABLE 4
Maximum Likelihood Estimation for Model 1
dep var./indep. var. Standardized
coefficients
Non
standard.
coefficients
Standard
error t value
man_opp/socializ -0,035 -0,035 0,046 -0,760
man_opp/size -0,160 -0,367 0,102 -3,592
man_opp/age -0,115 -0,164 0,068 -2,416
man_opp/hf -0,517 -1,644 0,153 -10,739
man_opp/hum_spec -0,312 -0,318 0,044 -7,263
man_opp/prop_uni -0,200 -0,008 0,002 -3,938
del_disc/man_opp -0,238 -0,277 0,062 -4,445
motiv/man_opp -0,341 -0,341 0,049 -6,938
motiv/del_disc 0,298 0,257 0,042 6,077
perform/motiv 0,297 3,756 0,695 5,401
perform/del_disc 0,211 2,298 0,562 4,092
perform/man_opp -0,158 -1,994 0,662 -3,010
ABSOLUTE GOODNES OF FIT
Chi-square = 304,171 / Probability level = 0,000
Goodness of Fit (GFI) = 0,883
Root Mean Square Error of Approximation (RMSEA) = 0,185
INCREMENTAL GOODNES OF FIT
Adjusted Goodness of Fit Index (AGFI) = 0,744
Tucker Lewis Index (TLI) = 0,444
Normed Fit Index (NFI) = 0,679
PARSIMONIOUS GOODNES OF FIT
Akaike Information Criterion (AIC) = 364,171
Parsimonious Normed Fit Index (PNFI) = 0,377
Parsimonious Goodness of Fit Index (PGFI) = 0,402

35
TABLE 5
Maximum Likelihood Estimation for Model 2
dep var./indep. var. Standardized
coefficients
Non
standard.
coefficients
Standard
error t value
man_opp/size -0.161 -0,368 0,102 -3,596
man_opp/age -0.113 -0,159 0,068 -2,347
man_opp/hf -0.533 -1,686 0,143 -11,802
man_opp/hum_spec -0.321 -0,326 0,042 -7,689
man_opp/prop_uni -0.202 -0,008 0,002 -3,947
del_disc/man_opp -0,277 0,063 -4,428
motiv/man_opp -0.342 -0,341 0,049 -6,913
motiv/del_disc 0.3 0,257 0,042 6,077
perform/man_opp -0.158 -1,994 0,665 -3,001
perform/motiv 0.297 3,756 0,695 5,401
perform/del_disc 0.212 2,298 0,562 4,092
ABSOLUTE GOODNES OF FIT
Chi-square = 52,735 / Probability level = 0,00
Goodness of Fit (GFI) = 0,967
Root Mean Square Error of Approximation (RMSEA) = 0,071
INCREMENTAL GOODNES OF FIT
Adjusted Goodness of Fit Index (AGFI) = 0,927
Tucker Lewis Index (TLI) = 0,900
Normed Fit Index (NFI) = 0,916
PARSIMONIOUS GOODNES OF FIT
Akaike Information Criterion (AIC) = 102,735
Parsimonious Normed Fit Index (PNFI) = 0,509
Parsimonious Goodness of Fit Index (PGFI) = 0,430