Description
An important question in the control literature concerns the role of interpersonal trust in
the design and functioning of formal control systems for collaborative settings. In this
paper, we argue that subordinate’s trust in the superior depends on the formality of the
performance evaluation procedure and that this relationship is mediated by managerial
perceptions of justice and feedback. The findings confirm our predictions. Furthermore,
we find that formality matters more for trust formation for those managers that are in
functions with less contractible outputs.
How formal performance evaluation affects trust between superior
and subordinate managers
Frank Hartmann
a,1
, Sergeja Slapnic?ar
b,
*
a
RSM Erasmus University, Rotterdam Burgemeester Oudlaan 50, 3062 PA Rotterdam, The Netherlands
b
University of Ljubljana, Faculty of Economics, Kardeljeva pl. 17, 1000 Ljubljana, Slovenia
a b s t r a c t
An important question in the control literature concerns the role of interpersonal trust in
the design and functioning of formal control systems for collaborative settings. In this
paper, we argue that subordinate’s trust in the superior depends on the formality of the
performance evaluation procedure and that this relationship is mediated by managerial
perceptions of justice and feedback. The ?ndings con?rm our predictions. Furthermore,
we ?nd that formality matters more for trust formation for those managers that are in
functions with less contractible outputs.
Ó 2008 Elsevier Ltd. All rights reserved.
Introduction
In contemporary business environments, interpersonal
trust is considered an important contributor to organiza-
tional performance (e.g., Dirks & Ferrin, 2002; Zaheer,
McEvily, & Perrone, 1998). Interpersonal trust is important
whenever organizational participants engage in situations
that demand collaborative effort (cf. Coletti, Sedatole, &
Towry, 2005). It is especially important in supervisor–sub-
ordinate relationships, as trust improves cooperation and
mitigates agency problems (e.g., Jones, 1995) by enhancing
information exchange and reducing subordinates’ per-
ceived need to engage in short term opportunistic behavior
(e.g., Fisher, Maines, Peffer, & Sprinkle, 2005). Although
important, interpersonal trust is generally not deemed
suf?cient for supervisors to control their subordinates’
behaviors. Therefore, supervisors typically use formal
controls such as performance evaluation systems (cf. Das
& Teng, 1998; Malhotra & Munighan, 2002). In this paper,
we investigate whether and how superiors’ use of perfor-
mance evaluation systems affects subordinates’ trust in
their superiors.
This investigation is warranted since the relationship
between performance evaluation and interpersonal trust
is not well understood. For example, Lau and Buckland
(2001) and Lau and Shohilin (2005) both investigated the
trust effects of superiors’ choice of ?nancial or non-?nan-
cial performance metrics for evaluating subordinate per-
formance, but report almost opposite ?ndings. Whereas
the former study concludes that trust is higher when supe-
riors use ?nancial performance measures, which are more
‘objective’ and ‘truthful’, the latter study reports that trust
is higher when superiors use non-?nancial measures,
which are ‘broader’ and ‘more complete’. Research in team
settings also provides mixed evidence on the effects of for-
mal controls on trust. Studies on trust attributions (e.g.,
Malhotra & Munighan, 2002; Tenbrunsel & Messick,
1999) suggest that formal control systems may decrease
interpersonal trust as they cause collaborative behavior
to be attributed to the presence of formal controls, rather
than to team member trustworthiness. In contrast, Coletti
et al. (2005) ?nd positive trust effects of using formal per-
formance evaluation systems in teams. They argue that
these systems are essential in providing collaborators with
performance feedback, which contain implicit signals
about mutual trustworthiness.
In combination, available evidence suggests that perfor-
mance evaluation systems have a potentially important
impact on trust, but also raises questions about which
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.11.004
* Corresponding author. Tel.: +386 1 5892 483; fax: +386 1 5892 698.
E-mail addresses: [email protected] (F. Hartmann), Sergeja.slapni-
[email protected] (S. Slapnic?ar).
1
Tel.: +31 (0) 10 40 82 222; fax: +31 (0) 10 40 89 017.
Accounting, Organizations and Society 34 (2009) 722–737
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
aspects of performance evaluation systems affect trust and
how, and about how the attribution effects that such sys-
tems have may affect the creation of trust.
In this paper, we address these questions by developing
and testing a path model that relates superior’s use of the
performance evaluation system to subordinate’s subse-
quent trust in the superior. To reconcile opposite ?ndings
from studies that focused on superiors’ choice of metrics
(e.g., Lau & Buckland, 2001; Lau & Shohilin, 2005), we pro-
pose that trust is associated with superiors’ formal use of
the performance evaluation system. We then explore two
paths by which superior’s formal use of the performance
evaluation system may affect subordinate’s trust: (i)
through the enhanced perceived justice of the performance
evaluation process and (ii) through the enhanced perceived
quality of performance feedback. We ?nally investigate
whether the path-relationships are stable across manage-
rial functions, arguing that the attribution of trustworthy
behavior to the superior, rather than the performance eval-
uation system, depends on managerial job characteristics.
By developing and testing this model our contribution
to the literature is fourfold. First, we aim to show that trust
is affected by the degree of formality in the use of perfor-
mance measures rather than the type of measures, thus
reconciling the opposite results from studies that looked
at the choice of performance metrics (Lau & Buckland,
2001; Lau & Shohilin, 2005). Second, by exploring two
mediators of the relationship between performance evalu-
ation system use and trust, we provide an explicit test of
the intervening mechanisms that previous studies merely
suggested but never addressed empirically. Third, we ex-
plain differences in trust attributions by differences in
managerial job characteristics in an attempt to reconcile
related ?ndings in previous studies (Coletti et al., 2005;
Malhotra & Munighan, 2002). Fourth and ?nally, this paper
integrates the literature on control and trust attributions in
team settings (e.g., Coletti et al., 2005) with the literature
on the hierarchical setting of superiors and subordinates,
the latter of which is an area of primary concern to the
management accounting literature at large (Hartmann,
2000, 2007).
The remainder of this paper is structured as follows.
Section ‘‘Literature review” provides an overview of the ex-
tant literature on the relationships between trust and per-
formance evaluation systems and identi?es some of the
gaps and con?icts in extant studies. Section ‘‘Hypotheses
development” develops the path model from which
hypotheses are derived about the relationship between
supervisory use of the performance evaluation system
and trust. Section ‘‘Method” presents the design of the
empirical survey that we conducted. Section ‘‘Results” pre-
sents the results from the tests of the hypotheses. Section
‘‘Discussion and conclusions” concludes this paper with
an overview of ?ndings, the study’s limitations and some
directions for future research.
Literature review
Despite the large variety of de?nitions and descriptions
of trust in organizational settings, both the organizational
and accounting literatures generally recognize the impor-
tance of trust for organizational functioning (cf. Coletti
et al., 2005; Rousseau, Sitkin, Byrt, & Camerer, 1998). In
accounting studies, trust appears to be an important factor
across organizational settings and levels of analysis, and
explains the functioning of company-wide pay-for-perfor-
mance systems as well as the success of inter-organiza-
tional networks (e.g., Chenhall & Lang?eld-Smith, 2003;
Dekker, 2004; Free, 2008; Tomkins, 2001). In this study,
we focus on the interpersonal trust of subordinates’ in
their superiors, which has received some attention in the
accounting literature over the last decades (e.g., Hopwood,
1972; Lau & Tan, 2006). We de?ne trust in this setting as
the belief that subordinates hold about the trustworthiness
of their superiors (cf. Clark & Payne, 1997; Coletti et al.,
2005). In accordance with the psychological view on trust,
we consider a person’s trustworthiness as that person’s
preference for upholding social and behavioral norms, such
as honesty and fairness, in situations in which the trusting
party will bene?t from the adherence to those norms (cf.
Rousseau et al., 1998).
2
The few studies that investigated the effects of supe-
rior’s performance evaluation behavior on subordinate’s
trust thus far seem to con?rm that upholding social norms
of honesty and fairness enhances trust, but yield con?ict-
ing results concerning the role of the performance evalua-
tion system (e.g., Hopwood, 1972; Lau & Buckland, 2001;
Lau & Shohilin, 2005; Ross, 1994). These studies have
mainly focused on the superior’s choice of performance
metrics for evaluating and rewarding subordinate perfor-
mance. Hopwood (1972), for example, found that subordi-
nates reported higher trust in their superiors, when these
superiors measured subordinates’ performance strictly
against budget. This ?nding was later con?rmed by Lau
and Buckland (2001), who showed that trust mediates
the relationship between the use of budget-based perfor-
mance metrics and job-related tension. In both studies,
these ?ndings were explained by the fact that subordinates
perceive the use of budget-related performance criteria as
the superior’s attempt to provide ‘precise’ and ‘honest’ per-
formance feedback. In contrast, in these studies the use of
non-?nancial performance metrics was associated with
lower levels of trust. The authors argued that their use
was perceived as ‘subjective’ and ‘inaccurate’, which is sur-
prising as no a priori reason exist why ?nancial metrics
would be more subjective and accurate than other (non-
?nancial) quantitative metrics.
Although these results seem to generally support the
importance of upholding norms of honesty and clarity for
superiors’ trustworthiness (cf. Clark & Payne, 1997;
2
The economic literature considers trust as the outcome of a calculation-
based prediction a trusting party makes about a trusted party’s future
behavior (see, e.g., Hardin, 1992; Kramer, 1999; Williamson, 1993). The
psychological notion of trust adopted here considers trust as the outcome
of affective and emotional considerations that form the trusting party’s
belief about the trusted party’s behavioral dispositions (cf. Fine & Holy?eld,
1996; March & Olsen, 1989). The psychological view on trust disputes an
important role of calculation in trust creation, as it considers the associated
calculations too complex for the cognitive capacities of any trusting party
(cf. Fine & Holy?eld, 1996).
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 723
Rousseau et al., 1998), further evidence on the relationship
between performance evaluation system use and trust
raises doubt about the conclusiveness of these ?ndings.
First, the results of Hopwood (1972) and Lau and Buckland
(2001) seem contradicted by the study of Lau and Shohilin
(2005). Lau and Shohilin (2005) followed a similar line of
reasoning, but found in a different sample that trust was
higher when superiors used non-?nancial performance
metrics. Their explanation was that non-?nancial metrics
have a higher accuracy as they enable a ‘broader’ perfor-
mance assessment in the eyes of the subordinates than
the use of ‘narrow and rigid’ budget-based metrics (Lau &
Shohilin, 2005, p. 396). Thus, although it seems undisputed
that upholding social norms may evoke trust in evaluative
settings, these ?ndings suggest that perceived adherence
to such norms is not related to the use of a particular kind
of metrics, and that different aspects of performance eval-
uation system use may be involved.
Second, studies on the attribution effects of control sys-
tems show that upholding norms of honesty and clarity by
a superior does not necessarily result in perceptions of
higher trustworthiness (e.g., Malhotra & Munighan, 2002;
Tenbrunsel & Messick, 1999). Indeed, when honest and
clear behavior is perceived to be caused by an external
control system, rather than a person’s inherent intentions,
trust will not emerge as such behavior is not attributed to
trustworthiness (Gambetta, 1988; Malhotra & Munighan,
2002). This is a central claim of attribution theory, which
is generally concerned with the question as to whether a
person explains the behavior of another person by situa-
tional factors or by the other person’s inherent dispositions
(cf. Birnberg, Frieze, & Shields, 1977; Gambetta, 1988). This
theory predicts that upholding social norms will only re-
sult in trust attributions when the person makes a costly
choice by upholding such norms, as this signals that the
person’s upholding of norms is intentional (cf. Gambetta,
1988; Malhotra & Munighan, 2002, p. 537). In their study
on team collaboration with and without formal controls,
Malhotra and Munighan (2002) thus expected and found
that using binding (i.e., formal) team contracts to control
team cooperation led to lower intra-team trust than the
use of non-binding (i.e., informal) team contracts. Binding
contracts caused team members to attribute cooperation
to the formal contract, while a voluntary adherence to
the non-binding contract was seen as a signal of inherent
trustworthiness. In their study on performance evaluation
systems in a team setting, Coletti et al. (2005) demon-
strated that such negative attribution effects may be com-
pensated by positive attribution effects of performance
evaluation system use on trust. They found that interper-
sonal trust increased when the performance evaluation
system enhanced cooperative performance, and when the
system provided team members with feedback on their
joint performance.
Overall, empirical evidence on performance evaluation
and trust strongly suggests that they are related, yet fails
to provide conclusive evidence on the aspect of the perfor-
mance evaluation system that matters and on the attribu-
tions that these systems may cause. In section ‘‘Hypotheses
development”, we will explore these two issues by devel-
oping set of related empirical hypotheses about the rela-
tionship between the use of performance evaluation
systems and trust.
Hypotheses development
The opposite ?ndings of studies that investigated the ef-
fect of performance metrics on trust (e.g., Lau & Buckland,
2001; Lau & Shohilin, 2005) can be reconciled by acknowl-
edging that in these studies trust may have been the con-
sequence of the performance measurement process,
rather than the type of performance metric used. The per-
formance evaluation literature indeed con?rms that supe-
riors differ in the extent to which they use available
systems and choose to apply explicit or implicit measure-
ment (Hartmann, 2000; Moers, 2005). This choice may en-
hance trust, as it signals the superior’s intention to uphold
norms of honesty and clarity, which only explicit measures
can provide (cf. Lau & Buckland, 2001; Lau & Shohilin,
2005). However, since performance evaluation is broader
than performance measurement alone, subordinates’ trust
responses are likely to be a reaction to all stages of the
performance evaluation process (Dirks & Ferrin, 2002).
Thus, to explain trust reactions to performance evaluation
system use, our analysis will also consider the choices that
superiors make concerning the development of perfor-
mance targets and the way in which superiors assign
performance related rewards (cf. Fisher et al., 2005;
Greenberg, 1986; Ittner & Larcker, 2001; Otley, 1999;
Zimmerman, 2003).
In line with the de?nitions of formal and informal con-
trol in the organizational literature (see, e.g., Morand,
1995; Sitkin & George, 2005), we propose to address these
choices in terms of the level of formality that superiors ap-
ply when using the performance evaluation system.
3
For
the different stages of the performance evaluation process
formality has the following meaning. Regarding target set-
ting, a high level of formality denotes the situation in which
superiors explicate performance targets in quantitative and
written terms. Informal target setting, instead, means having
implicit, qualitative (‘do your best’) targets, that are commu-
nicated informally, and whose subsequent achievement can-
not be measured objectively (cf. Locke & Latham, 1990).
Regarding performance measurement, a high level of for-
mality implies the use of quantitative and objective, rather
than qualitative and subjective measures, which make up
informal performance assessments (cf. Moers, 2005).
4
3
In the wider organizational literature the terms formal control and
informal control have various connotations beyond our strict de?nition.
‘Informal control’ is seen as ‘loose’, ‘spontaneous’ and ‘casual’. ‘Formal
control’ instead re?ects as ‘tight’, ‘deliberate’ and ‘bureaucratic’ (see, e.g.,
Morand, 1995, p. 831).
4
Moers’ (2005, p. 71) provides little theoretical justi?cation for his
distinction between ‘objective’ and ‘subjective’ measures. Therefore, it is
not clear whether the distinction relates to the measure’s properties, to the
process by which they are derived, nor to whose accord the measure should
be seen as ‘objective’ or ‘subjective’. According to the examples given in the
paper, ‘objective measures’ seem to be those that include some explicit
numerical standard (e.g., ‘2% cost decrease’), whereas ‘subjective’ measures
simply refer to activities (e.g., good use of resources’). We believe that this
distinction corresponds with our more precise distinction between high
and low levels of formality.
724 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
Regarding the assignment of performance-based rewards, a
high level of formality means the formulaic determination
of rewards and bonuses, rather than the use of untraceable
personal judgment (cf. Gibbs, Merchant, Van der Stede, &
Vargus, 2004). In sum, superiors that use the performance
evaluation in a formal way are those superiors that explicate
performance targets, measure performance by a set of clear
metrics, and who give rewards based on clear allocation
rules. We regard these three phases of the performance eval-
uation process as three-dimensions of the performance eval-
uation system that determine its overall formality (see
Fig. 1).
We expect that a formal use of the performance evalu-
ation system, as de?ned above, will have a positive effect
on trust, as it allows higher levels of integrity, honesty,
accuracy and consistency in performance evaluation than
informal performance evaluation. This is not only consis-
tent with the studies by Hopwood (1972), Lau and
Buckland (2001) and Lau and Shohilin (2005), but also with
other recent evidence on the behavioral effects of implicit
and subjective performance measures (e.g., Hartmann,
2007; Moers, 2005). Moers (2005), for example, recently
shows that subjective performance evaluations result in
biases, such as ‘compression’ (i.e., low-rating variance),
and ‘leniency’ (i.e., high-rating average), which are clearly
at odds with the social norms of honesty, accuracy and
consistency that form the necessary conditions for subor-
dinates’ trust in their superior (e.g., Rousseau et al., 1998;
Whitener, Brodt, Korsgaard, & Werner, 1998). In their
study on the trust effects of managerial decision-making,
Sitkin and George (2005) show that when superiors have
to make painful or controversial personnel decisions they
tend to counter threats to their trust and legitimacy by
increasing their use of formal controls and reducing their
use of informal controls.
Before exploring more detailed relationships between
performance evaluation system use and trust, we therefore
propose to test the following relationship as an overall
starting point of our analysis:
H1: Formal use of the performance evaluation system
by superiors has a positive effect on subordinates’ trust
in superiors.
Formality, trust and the mediating effect of procedural justice
Trust is a consequence of managers’ adhering to social
norms as integrity, honesty, accuracy and consistency (cf.
Clark & Payne, 1997, p. 208; Whitener et al., 1998). We ex-
pect that these factors mediate the relationship between
performance evaluation and trust (H1) because they deter-
mine the procedural justice of the performance evaluation
process.
The organizational justice literature demonstrates that
the social norms that evoke trust show great resem-
blance to the social norms that determine whether orga-
nizational decision-making processes will be judged ‘fair’
by the stakeholders of these processes (e.g., Leventhal,
1976; Leventhal, Karuza, & Fry, 1980). Leventhal (1976,
p. 39), for example, argues that decision-makers should
adhere to such social norms as consistency, freedom
from bias, and accuracy to enhance the procedural
justice of the decision-making processes they engage in.
Therefore, when superiors uphold the social norms of
honesty, accuracy and consistency, procedural justice is
enhanced and trust is likely to emerge (e.g., Rousseau
et al., 1998; Whitener et al., 1998). In their meta-analytic
review Dirks and Ferrin (2002) con?rm the positive rela-
tionship between procedural justice and trust in extant
research.
Since we expect that procedural justice mediates the
relationship between performance evaluation and trust,
we propose that procedural justice is a direct and positive
consequence of superiors’ formal rather than informal use
of the performance evaluation system when making per-
formance evaluation decisions. Superiors who explicate
performance targets, measure performance by clear met-
rics, and give rewards based on clear allocation rules (i.e.,
high formality) will be able to deliver more consistent,
more accurate and less biased performance evaluations
(i.e., higher procedural justice) than superiors who use
implicit targets, measure performance subjectively and
allocate rewards using mere personal judgment. This is
consistent with the limited and partial evidence from the
accounting literature (e.g., Hopwood, 1972; Lau &
Buckland, 2001), as well as with ?ndings in some early
organizational studies that documented the fairness conse-
formality trust
FormTS
FormP
Formrew
Procj1 Procj2 Procj3
Feed1 Feed2 Feed3
Trust1
Trust2
Trust3
Procj3
Feed3
feedback
quality
procedural
justice
uncertainty
tenure
Fig. 1. Full model (all predicted relationships are positive).
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 725
quences of performance evaluation process characteristics
(e.g., Greenberg, 1986; Landy, Barnes, & Murphy, 1978).
5
Colquitt and Jackson (2006, p. 871) recently found that fair-
ness criteria should apply to all stages of a decision-making
process, such as the information search, the choice of deci-
sion rules, and the communication of outcomes. This sup-
ports our expectation that trust results from superiors
acting consistently, accurately and honestly throughout the
whole performance evaluation process, rather than giving
subordinates a favorable evaluation as such (Taylor, Master-
son, Renard, & Tracy, 1998; Taylor, Tracy, Renard, Harrison,
& Carroll, 1995).
Overall, therefore, we expect that when superiors expli-
cate performance targets in formal terms, attempt to quan-
tify measurement and allocate rewards using formulaic
and traceable procedures, this enhances procedural justice
and therefore trust. This is expressed in the following
hypothesis:
H2: The positive effect of the formal use of the perfor-
mance evaluation system on trust is mediated by per-
ceived procedural justice, such that the formal use of
the performance evaluation system has a positive effect
on perceived procedural justice, which in turn has a
positive effect on trust.
Formality, trust and the mediating effect of performance
feedback quality
Our review of accounting studies on trust antecedents
suggests a second mechanism by which performance eval-
uation procedures affect trust, which is via the perfor-
mance feedback that such evaluation systems typically
provide (e.g., Coletti et al., 2005; Lau & Buckland, 2001). In-
deed, these studies suggest that the effect of performance
evaluation on trust may be mediated by the increased
quality of performance feedback that subordinates obtain,
in cases when superiors evaluate performance formally,
rather than informally. Performance feedback is de?ned
as the feedback that subordinates get about the perfor-
mance levels they attain (cf. Steelman, Levy, & Snell,
2004). Such feedback is considered an essential function
of any performance evaluation process, as it should help
subordinates understand how they may adapt their behav-
iors to enhance performance (e.g., Herold & Parsons, 1985;
Steelman et al., 2004), and thus to improve collaboration
with their superior. Last, but not least, good feedback al-
lows subordinates to adjust the expectations about their
rewards (Cahuc & Zylberberg, 2004). Although many fac-
tors affect feedback quality (cf. Herold & Parsons, 1985),
the management literature strongly suggests that effective
feedback on individual performance requires that targets
are set clearly (e.g., Hartmann, 2007; Locke & Latham,
1990), that performance is measured accurately and in
an understandable way (Steelman et al., 2004), and that
reward distribution is transparent and traceable (cf. Moers,
2005). Formal evaluations have an advantage over infor-
mal evaluations in that they clearly specify the dimensions
of performance that are being evaluated, and the way in
which the rewards are linked to those performance dimen-
sions (cf. Marginson & Ogden, 2005; Moers, 2005), which
enhances goal orientation and motivation (cf. Hartmann,
2005; Hartmann, 2007). Thus, we predict that formal eval-
uations are positively related to trust because they provide
subordinates with a higher quality of feedback than do
informal evaluations (cf. Lau & Buckland, 2001; Moers,
2005). Coletti et al. (2005) show that receiving high-qual-
ity feedback will positively affect trust, as subordinates
will perceive superiors’ attempts to provide good feedback
as an indication of superiors’ desire to help subordinates
improve their performance as well as to improve their
working relations. These factors are all part of the social
norms underlying trustworthiness (Rousseau et al.,
1998). Consistent with extant evidence on the relation-
ships between performance evaluation, feedback and trust,
we therefore predict that the overall positive effect of for-
mality on trust is mediated by the quality of performance
feedback as well. This is summarized in the following
hypothesis:
H3: The positive effect of the formal use of the perfor-
mance evaluation system on trust is mediated by the
perceived quality of performance feedback, such that
the formal use of the performance evaluation system
has a positive effect on perceived quality of perfor-
mance feedback, which in turn has a positive effect on
trust.
It is clear that the theory underlying hypotheses 1, 2
and 3 suggests relatedness between feedback quality and
procedural justice, since the underlying characteristics
share several similarities. The factors such as accuracy
and lack of bias seem equally important for evoking per-
ceptions of procedural justice, as well as for providing
high-quality feedback. In large parts of the organizational
justice literature, this relatedness has been recognized,
such that performance feedback itself is considered an
important antecedent of procedural justice (e.g., Cohen-
Charash & Spector, 2003; Taylor et al., 1995). Thus,
according to these studies on procedural justice, good
performance feedback is a constitutive element of a fair
(evaluation) process, as it provides the openness and high
level of communication (e.g., Colquitt & Jackson, 2006;
Erdogan, 2002) that fair processes demand. We therefore
expect that perceived quality of performance feedback
not only has a direct effect on trust, as formulated in
hypothesis 2, but also affects trust via procedural justice.
We therefore propose to test the following hypothesis as
part of the overall model:
H4: The perceived quality of performance feedback
has a positive impact on perceived procedural
justice.
5
The survey study by Landy et al. (1978) reports positive correlations
between ‘fairness’ of performance evaluation and the existence of a formal
evaluation program, and formal goal setting. The survey study by Green-
berg (1986) reports ‘consistent application of standards’ as an important
correlate of procedural justice of performance appraisals.
726 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
Formality of performance evaluation, trust and managerial
job characteristics
Our expectations so far implicitly assume that when
subordinates receive fair and high-quality performance
evaluations, at least some of these qualities will be attrib-
uted to the superior providing the evaluations. This is con-
sistent with available empirical evidence (e.g., Coletti et al.,
2005; Hopwood, 1972; Lau & Buckland, 2001). However, to
explicitly account for attribution effects that formal control
systems may cause, we now explore whether the relation-
ships identi?ed are likely to be affected by such
attributions.
Generally, attribution theory suggest that the emer-
gence of trust requires that subordinates perceive superi-
ors’ provision of fair and high-quality performance
evaluations as a signal of their inherent trustworthiness,
rather than as a consequence of the formal control system
in place (cf. Malhotra & Munighan, 2002). In other words,
although subordinates may be generally inclined to trust
superiors who provide honest, accurate and consistent per-
formance evaluation and clear performance feedback, the
effect on trust will be higher if subordinates believe that
these behaviors are a consequence of superior’s deliberate
and costly intentions.
As regards the positive effects of the formal use of the
performance evaluation system, expressed in hypotheses
1–3, we expect that the extent to which the formal use
of the performance evaluation system results in trust, via
perceptions of higher fairness and higher quality of perfor-
mance feedback, may depend on managerial job character-
istics. Based on attribution theory, we expect that the
extent to which subordinates will attribute superior’s for-
mal use of the performance evaluation to their inherent
trustworthiness depends on the extent to which formal
performance evaluation is costly to the superior, that is,
when it requires high levels of effort of the superior (cf.
Gambetta, 1988; Herold & Parsons, 1985; Malhotra &
Munighan, 2002). We argue that such will be the case for
those subordinates’ functions for which outputs are not
easily measurable in quantitative and objective terms.
For brevity purposes, we use the term contractibility to
capture the ability to measure outputs (cf. Grossman &
Hart, 1986). Examples of high-contractibility outputs
include typical ?nancial and quantitative indicators of
turnover and pro?t, and other technical outputs for which
measures and targets are relatively easy to develop
(cf. Moers, 2005). Such outputs are widely available in
organizational functions where subordinates are engaged
in functions related to external trade and sales. Examples
of low-contractibility outputs include those in which
measurability of individual performance is hindered by
organizational interdependencies, and in which job tech-
nology causes a lack of objective metrics. This refers to
functions in which subordinates are responsible for provid-
ing internal service, advice and support (cf. Moers, 2005).
In jobs for which outputs are not easily contractible, supe-
riors have to apply extra effort to explicate performance
targets, measure performance clearly, and give rewards
based on clear allocation rules than in jobs with easily
contractible outputs (Hartmann, 2007). For our model this
implies the following.
Regarding procedural justice, we expect that subordi-
nates in functions with relatively contractible outputs
(e.g., in sales functions) will be less prone to attribute per-
formance evaluation formality to superiors’ inherent trust-
worthiness. As superiors can use easily available metrics
for target setting, performance measurement and reward
distribution for these functions, formality does not provide
a strong signal to subordinates about superiors’ true and
costly intentions toward providing fair evaluations. Thus,
formality will add less to fairness attributions in such func-
tions, than in functions in which formality requires more
effort on the part of the superior (cf. Sitkin & George,
2005). This situational dependency of fairness judgment
is consistent with recent ?ndings by Colquitt and Jackson
(2006), who showed that fairness judgment may indeed
depend on context.
Regarding the second mediating variable, performance
feedback, a similar argument applies. In situations where
subordinates already have good understanding of their
own performance levels and the consequences of this per-
formance level for individual wealth, the bene?ts of formal
performance evaluation in terms of providing feedback
information will be less prevalent. Here, attribution theory
suggests that the outcomes of formal ways of performance
evaluation are likely to be considered the consequence of
the formal system, rather than the outcome of supervisor’s
deliberate effort to provide such information (cf. Sitkin &
George, 2005). In contrast, subordinates in functions with
less contractible outputs will thus be more inclined to
attribute the formal use of the performance evaluation sys-
tem to their supervisor’s inherent and costly inclination to
provide accurate performance feedback.
In sum, we expect that the relationship between the
formal use of the performance evaluation system and trust
is different for different levels of contractibility. We expect
that contractibility affects the extent to which feedback
quality and justice are considered signs of trustworthiness.
In particular, our expectation can be summarized as
follows:
H5: The positive effect of the formal use of the perfor-
mance evaluation system on perceived performance
feedback quality and justice, and consequently on trust,
will be stronger for managers in functions with less
contractible outputs.
The proposed hypotheses are summarized in the model
in Fig. 1.
Method
To test these hypotheses, we conducted a survey study
among 160 departmental managers in eleven Slovenian
commercial banks. The choice of a survey study is based
on two considerations. First, there are no publicly available
archival data on the constructs that matter for our study.
Second, studies into justice and trust are usually perceived
to concern private information, which requires that the
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 727
data collection be anonymous. This is easily achieved by
the survey method. The choice for the banking sector
was based on the homogeneity of the sector, the compara-
bility of ?rms and, most importantly, the variation in con-
tractibility of outputs between managers.
To select survey respondents, we approached HR-man-
agers of all 18 Slovenian banks at the board level through
university contacts. Boards of directors of 11 banks agreed
to participate in the research. With the help of the human
resource board members, respondents in each bank were
selected based on the following three criteria. Managers
had to have at least one subordinate. They had to be in
their current position for at least one year to guarantee
their familiarity with the performance evaluation system.
Their compensation had to be contingent on performance,
and they had to experience at least one evaluation cycle, to
make sure that the evaluation system was a salient feature
of their working environment. Following these criteria a
total of 260 managers were selected for participation by
the banks’ HR-departments. There were a total of 160
usable responses, which represents a 61.5% response rate.
Table 3 reports on the number of managers for each of
the ?rms.
The survey instrument was web-based. We obtained e-
mail addresses of all respondents, and sent an individual-
ized e-mail message to each respondent with the request
to participate, a reference to the endorsement from the
bank, a unique and anonymous code, and a link to a web-
page of one of the sponsoring universities. Respondents
could login to the web-based questionnaire using the code.
In the design of the survey instrument, available guidelines
were followed (Dillman, 2000). Several measures were ta-
ken to increase the overall quality of the instrument, as
well as the response rate. First, the website was designed
by an experienced designer, and was made attractive by
using a limited number of colors, and an overall attractive
layout. Second, when possible, we used items to measure
our constructs from the extant literature. Third, the full
questionnaire was drafted in English, after which it was
translated into Slovenian using a translate–retranslate pro-
cedure. Four, the Slovenian questionnaire was subse-
quently tested with ?ve bank managers from three
different banks to assess the length, the understandability,
the attractiveness, and to search for any further dif?culties
respondents might have. In this phase small adaptations
were made to the wording of some questionnaire items
and the structure of the web instrument. This version of
the questionnaire was also intensively discussed with the
HR-managers. Finally, after sending the e-mails, a follow-
up procedure was used in which respondents were re-
minded two times by their employing organization based
on the codes provided by the researchers. Tables 1–3 con-
tain descriptive information on the sample of respondents.
The number of observations per organization ranged
from 4 to 35, re?ecting the organization’s size and respon-
sibility center structure (see Table 3).
Variable measurement
We measure formality of the performance evaluation
system (FORM) as a latent construct, for which we used
three manifest variables re?ecting each of the three steps
in the performance evaluation cycle (i.e., target setting,
performance measurement and rewarding) separately (cf.
Ittner & Larcker, 2001; Otley, 1999). We developed the
measurement instrument for formality of different subsys-
tems as follows.
Formality of target setting (FormTS) is measured using
four potential dimensions of performance in accordance
with the balanced scorecard logic used in Ittner, Larcker,
and Meyer (2003, see also the Appendix). We assessed
the relevance of these four-dimensions for our sample of
bank managers, as well as the examples of the perfor-
mance metrics in each category during the interviews in
the pre-test phase. We accounted for potential variation
of formality across dimensions and for the relative impor-
tance of each dimension by adopting the following mea-
surement procedure. First, we measured on ?ve-point
Likert scales (i) the extent to which performance targets
are explicitly documented in a written form and (ii) the ex-
Table 1
Sample descriptives.
Mean
(median)
Standard
deviation
Range Valid n
Age 43.45 (44.00) 8.18 29–58 153
Years in bank 9.82 (8.58) 8.58 1–35 157
Tenure in function 4.76 (3.25) 4.73 1–30 158
Table 2
Respondents’ areas of responsibility.
Job Subgroup Frequency Percent Cumulative
Trading Front-of?ce 59 36.6 38.1
Back-of?ce Back-of?ce 43 26.7 65.8
Investment banking
and asset
management
Front-of?ce 8 5.0 71.0
Treasury Front-of?ce 4 2.5 73.5
Administration Back-of?ce 35 21.7 96.1
International operations Front-of?ce 4 2.5 98.7
Other Back-of?ce 2 1.2 100.0
Missing n.a. 5 3.7
Total 160
Table 3
Participating banks and frequency of respondents.
Bank Frequency Percent Cumulative
Bank 1 13 8.1
Bank 2 21 13.1 8.1
Bank 3 35 21.9 21.3
Bank 4 15 9.4 43.1
Bank 5 11 6.9 52.5
Bank 6 13 8.1 59.4
Bank 7 9 5.6 67.5
Bank 8 14 8.8 73.1
Bank 9 20 12.5 81.9
Bank 10 5 3.1 94.4
Bank 11 4 2.5 97.5
Total 160 100.0 100.0
728 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
tent to which targets are quanti?ed, for each dimension
separately. We also asked respondents to allocate 100
points to the four dimensions to re?ect the importance of
a certain performance dimension for the respondent’s area
of responsibility. These weights were used to obtain a ?nal
score for formality of target setting.
Formality of performance measurement (FormPM) is
measured in accordance with the procedures outlined for
formality of target setting. We measure on ?ve-point Likert
scales (i) the extent to which the superior relies on objec-
tive information from the information system and (ii) the
extent to which respondent’s evaluation is expressed in
quantitative terms. Scores are obtained for the four perfor-
mance dimensions, which are weighted as described
before.
Formality of rewarding (FormREW) is measured with
items that address the objectivity of reward determination,
using ?ve-point Likert scales. The questions address
whether rewards are based on the superior’s judgment,
as opposed to objective information from the information
system, and whether rewards are based on a quantitative
as opposed to qualitative link with performance. These
questions were answered for both the ?xed and the vari-
able part of managerial rewards.
The two mediating variables are measured as follows.
Feedback quality of the supervisor (FEED) measurement
is based on the list of indicators developed by Steelman
et al. (2004), for which we used ?ve-point Likert scales. A
sample item is ‘My supervisor gives me useful feedback
about my job performance’. Procedural justice (PROCJ) is
measured with four items that address the extent to which
respondents believe that the subsystems of target setting,
performance measurement and rewarding, as well as the
system in total, leads to a fair determination of pay. A sam-
ple item is ‘I have full con?dence in the system’s fairness in
evaluating performance’. This particular focus of the
questions re?ects that our study investigates subjective
perceptions of procedural justice, in contrast with the vast
majority of studies that address either the antecedents or
effects of procedural justice (cf. Cropenzano, Byrne,
Bobocel, & Rupp, 2001). Thus, our questions are not based
on any normative model of justice (cf. Leventhal et al.,
1980), to avoid testing a tautology.
6
Trust in supervisor (TRUST) is measured by the instru-
ment developed by Read (1962). It addresses trust by ask-
ing subordinates about the extent to which they perceive
that their superior is inclined to take actions that are in
the interest of the subordinate. These questions were se-
lected as they are neutral vis-à-vis speci?c causes of trust,
and directly assess elements of trustworthiness. A sample
item is ‘My superior will always act in my favor if he has
the chance’.
Contractibility of managerial outputs was proxied by
the function of the manager in the bank. We distinguished
between two groups of managers according to their job
characteristics. The ?rst group consists of managers with
jobs that involve transactions that have direct ?nancial
consequences for the bank, which are measurable and ver-
i?able. These managers are in charge of front-of?ce depart-
ments that grant loans, hire deposits and other forms of
loans, manage assets, do transactions on foreign markets
and similar. The second group of managers provide sup-
port to the ?rst group in terms of expertise, analysis, con-
trol and administration. These activities include the
analyses of projects, ful?lling reporting requirements,
internal audit, legal counselling and internal administra-
tion. These back-of?ce functions, while crucial for bank
performance, have less measurable and veri?able outputs,
and their effect on bank performance is less traceable.
Thus, for performance evaluation, this distinction seems
to capture the essential difference between managerial
performance of the ?rst group that is immediately and di-
rectly re?ected on the bank’s market performance, and the
performance effects of the second group’s activities which
are less obvious, and only indirectly and imperfectly linked
to market transactions.
We included two control variables which are tenure
(TENURE) and task uncertainty (UNCERT). The former is
measured with number of years the respondent has been
on the present managerial position. It controls for the trust
effects of tenure (cf. Gibbs et al., 2004). The latter is mea-
sured with the instrument developed by Withey, Daft,
and William (1983). This variable controls for uncertainty’s
potential effect on the use of performance measures.
Results
We tested our hypotheses by using a partial least
squares (PLS) method for structural equation model esti-
mation SEM (Lohmöller, 1989; Wold, 1982; Wold,
1985). Contrary to the variance–covariance-based method
of SEM estimation that attempts to minimize the differ-
ence between sample covariances and those predicted
by a theoretical model on the underlying assumptions
of normal multivariate distribution (Chin, 1998), PLS esti-
mates the structural model using an iterative OLS regres-
sion-like procedure, which aims to explain variance of the
dependent variables by minimizing the residual variance
of all dependent variables (both, latent and observed). Be-
cause of its iterative estimation algorithm that proceeds
block-by-block, PLS requires less stringent assumptions
about the distributional characteristics of the raw data
and sample size. We used SmartPLS software version
2.00.
Before the assessment of the structural model, we as-
sessed the quality of the measurement model, thus
addressing individual item reliability, construct reliability,
and convergent and discriminant validity for our re?ective
constructs (cf. Bagozzi, 1994). Individual item reliability is
considered adequate when an item has a factor loading
that is greater than 0.71 on its respective construct which
implies that more than 50% of the variance in the observed
variable is shared with the construct (cf. Chin, 1998). The
measure of internal consistency or construct reliability is
6
Several studies in the procedural justice literature apply a normative
model of justice that prescribes what process characteristics make a
process ‘fair’ (e.g., Taylor et al., 1998). Measuring the perceived presence of
these characteristics would be tautological as a procedural element would
then be an independent and a dependent variable at the same time.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 729
Dillon-Goldstein q (Tenenhaus, Vinzi, Chatelin, & Lauro,
2005). This measure is preferable to the Cronbach’s alpha,
as the latter assumes that all loadings are equal to one. A
threshold of 0.70 is suggested as a modest level of reliabil-
ity (Vandenbosch, 1996). Convergent and discriminant
validity of the constructs is assessed by the average vari-
ance extracted (AVE), which represents the average vari-
ance shared between a construct and its indicators
(Fornell & Larcker, 1981). For convergent validity, AVE
should be greater than 0.50. This ensures that measure-
ment error does not dominate the variance captured by
the construct (Vandenbosch, 1996). To assess discriminant
validity taking into account model parameters and mea-
surement errors, AVE should be greater than the variance
shared between any two constructs and the cross-loadings
of indicators on other constructs should be minimal.
The second step in the PLS analysis is the estimation of
the speci?ed structural equations. The path coef?cients
indicate the strength and direction of the relationships
among the latent variables. We assessed statistical signi?-
cance of parameter estimates using a bootstrap procedure
with 1000 replacements. In addition, the predictive valid-
ity of the parameter estimates can be assessed via a
cross-validated redundancy index or a so called Stone–
Geisser Q
2
-test (Geisser, 1974; Stone, 1974). A result indi-
cating that Q
2
for all latent variables is greater than 0 sug-
gests that the model has predictive relevance. As PLS
models lack an index that can provide the goodness of ?t
statistics like in variance–covariance-based SEM-ML mod-
els, Tenenhaus et al. (2005) and Vandenbosch (1996) argue
that next to the reliability and validity of constructs, the
signi?cance of variance explained and positive Q
2
s for all
constructs provide suf?cient evidence of model ?t.
The results of the measurement model for the full sam-
ple are summarized in Tables 4–6. As Table 4 reports, we
established that all factor loadings were greater than 0.71
with the exception of one indicator of TRUST. This provides
evidence of a satisfactory level of individual item reliabil-
ity. The indicator for TRUST, for which the loading of 0.69
was obtained, was maintained in the analysis given the
satisfactory construct reliability measured by q
c
, which ex-
ceeded 0.83 in all cases. Convergent validity appeared
acceptable for all re?ective constructs, as AVE exceeded
0.5 in all cases. Discriminant validity of constructs is pre-
sented in Table 5. It shows that diagonal AVE values exceed
all other scores, suggesting suf?cient discriminant
validity.
7
We performed our analysis of the hypotheses in a step-
wise fashion (Baron & Kenny, 1986; Luft & Shields, 2003),
such that we ?rst tested whether formality affected trust
directly (hypothesis 1, see Table 7, direct relationship mod-
el), and how this relationship altered by the subsequent
introduction of our two mediators according to hypothesis
2 and 3 (see Table 7, partially mediated relationship and
fully mediated relationship model). The results of the anal-
ysis for the full sample are presented in Table 7 panel A and
signi?cant (positive) paths are depicted as solid lines in
Fig. 2. The path coef?cient from FORM to TRUST shows
support for hypothesis 1, as FORM has a positive and statis-
tically signi?cant direct impact on TRUST (0.247, p < 0.001,
see direct relationship model). This effect is reduced when
PROCJ is introduced into the model, although the results do
not fully support hypothesis 2 as only the relationship be-
tween PROCJ and TRUST is signi?cant (0.20, p < 0.001),
whereas the direct path between FORM and PROCJ is insig-
ni?cant. When FEED is introduced into the model (hypoth-
esis 3) the direct link between FORM and TRUST is further
reduced and becomes insigni?cant, displaying a full medi-
ation effect (cf. Baron & Kenny, 1986). Both variables now
absorb the explanatory power of the initial direct effect of
FORM on TRUST. For FEED, the path coef?cient to TRUST is
0.412 (p < 0.001). The path coef?cient from FORM to FEED
is positive and signi?cant too (0.254, p < 0.01), which sup-
ports hypothesis 3. In support of hypothesis 4, FEED also
mediates the relationship between FORM and PROCJ as
the path coef?cient of FEED to PROCJ is 0.413 (p < 0.001).
Note that these results are obtained controlling for tenure
and task uncertainty.
Hypothesis 5 expressed our prediction that the relation-
ships between FORM and the two mediating variables
would be different for managers with more and less con-
tractible outputs. To test this hypothesis, we re-estimate
both the measurement model and the structural model
for the two subgroups of managers in front-of?ce and
back-of?ce functions, that we used as a proxy for high
and low contractibility of managerial outputs, respectively
(cf. Grossman & Hart, 1986). Table 2 shows how managers
in different areas of responsibility were allocated to the
subgroup analysis. Measurement model and reliability/
validity analyses for these subgroups are not presented
here for brevity purposes, but show no material differences
with the presented results for the full sample. The path
models for the front-of?ce subgroup and the back-of?ce
subgroup are presented in Table 7, panels B and C, respec-
tively. Figs. 3 and 4 depict the signi?cant positive paths for
the two subgroups.
The results suggest that for the high-contractibility sub-
group, formality of performance evaluation has no effect
on trust in the direct relationship model, nor is it linked
to procedural justice and feedback quality in the mediated
relationship model. For the low-contractibility subgroup,
the results are largely in line with the results obtained
for the full sample. Formality appears to affect trust (direct
relationship model), but the relationship appears to be
mediated by both procedural justice and feedback quality
(fully mediated relationship model). In the indirect model,
feedback quality also appears to mediate the relationship
between formality and procedural justice, as was expected.
In line with the results for the full sample, this suggests
that for the back-of?ce subgroup, feedback quality is the
crucial variable in the formality–trust relationship. As ex-
7
When more items per latent construct were available, we used
exploratory factor analysis using SPSS with Varimax rotation to identify
the best indicators for use in the measurement model. After eliminating
nine items on the basis of their low loadings on the respective factors (Hair,
Anderson, Tatham, & Black, 1995), we retained 14 indicators for the four
re?ective constructs (see the Appendix for an overview of the latent
constructs and their indicators). We performed a test of response bias by
comparing average scores on all variables for early and late respondents,
but found no evidence of response bias above chance (Blair & Zinkhan,
2006). The questionnaire is available from the corresponding author.
730 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
pected the same relationship is insigni?cant in the high-
contractibility subgroup. Such, too, is the relationship be-
tween procedural justice and trust in this subgroup. This
suggests that for the front-of?ce subgroup, formality of
performance evaluation does not invoke subordinate’s
trust of the superior, nor that higher formality would pos-
itively impact the perception of procedural justice or qual-
ity of feedback. As shown in Table 8, the differences in
strength of formality on feedback quality and perceptions
of justice between the two subgroups are signi?cant.
8
Discussion and conclusions
The purpose of this paper is to explore the effects of
superiors’ use of the performance evaluation system on
Table 4
Estimation of the measurement model parameters (full sample, n = 160).
Loading original sample Loading sample mean Standard deviation t-Statistic (bootstrap) Composite reliability
(q
c
)
Average variance
extracted (AVE)
FORM 0.882 0.715
FormTS 0.837 0.831 0.060 13.832
FormPM 0.793 0.785 0.074 10.730
FormRew 0.904 0.899 0.057 15.797
FEED 0.903 0.700
Feed
1
0.805 0.806 0.030 26.469
Feed
2
0.858 0.853 0.035 24.774
Feed
3
0.888 0.884 0.027 33.487
Feed
4
0.793 0.791 0.045 17.542
PROCJ 0.910 0.718
Procj
1
0.784 0.784 0.044 17.838
Procj
2
0.838 0.835 0.041 20.201
Procj
3
0.899 0.898 0.020 44.558
Procj
4
0.864 0.864 0.027 31.944
TRUST 0.845 0.647
Trust
1
0.847 0.846 0.028 30.219
Trust
2
0.871 0.869 0.025 34.501
Trust
3
0.681 0.676 0.080 8.507
UNCERT 0.902 0.754
Uncert
1
0.897 0.873 0.146 6.143
Uncert
2
0.882 0.856 0.149 5.903
Uncert
3
0.823 0.795 0.137 6.014
TENURE n.a. n.a.
Table 5
Discriminant validity coef?cients (full sample, n = 160).
FORM TENURE UNCERT PROCJ FEED TRUST
FORM 0.845
TENURE À0.140 n.a.
UNCERT 0.179 À0.008 0.868
PROCJ 0.172 À0.034 0.280 0.883
FEED 0.254 À0.160 0.163 0.430 0.836
TRUST 0.274 0.014 0.265 0.408 0.523 0.804
Diagonal elements (bold) are the square root of the variance shared
between the constructs and their indicators (AVE). Off-diagonal elements
are the correlations among constructs. For discriminant validity, diagonal
elements should be larger than off-diagonal elements. n.a. is non-
applicable.
Table 6
Cross-loadings (full sample, n = 160).
FORM FEED TRUST PROCJ UNCERT TENURE
FORM
FormPM 0.837 0.168 0.107 0.060 0.085 À0.117
FormTS 0.793 0.148 0.201 0.070 0.133 À0.034
FormRew 0.904 0.279 0.309 0.228 0.195 À0.170
FEED
Feed
1
0.226 0.805 0.480 0.448 0.152 À0.150
Feed
2
0.178 0.858 0.413 0.321 0.130 À0.113
Feed
3
0.207 0.888 0.443 0.365 0.127 À0.077
Feed
4
0.237 0.793 0.400 0.277 0.132 À0.200
TRUST
Trust
1
0.209 0.475 0.847 0.317 0.225 0.069
Trust
2
0.237 0.499 0.871 0.409 0.249 À0.024
Trust
3
0.232 0.218 0.681 0.229 0.147 À0.021
PROCJ
Procj
1
0.120 0.257 0.314 0.784 0.171 À0.005
Procj
2
0.157 0.464 0.342 0.838 0.250 À0.055
Procj
3
0.165 0.372 0.374 0.899 0.308 0.009
Procj
4
0.133 0.333 0.351 0.864 0.205 À0.059
UNCERT
Uncert
1
0.178 0.120 0.240 0.228 0.897 0.004
Uncert
2
0.167 0.196 0.283 0.315 0.882 À0.051
Uncert
3
0.106 0.097 0.141 0.166 0.823 0.041
TENURE
Tenure
1
À0.140 À0.160 0.014 À0.034 À0.009 1.000
8
We also analyzed the signi?cance of the difference in path-coef?cients
for the direct model between the two subgroups. These coef?cients were
0.026 and 0.370 for the front-of?ce and back-of?ce subgroups, respectively,
with the difference signi?cant at p < 0.05. This con?rms the outcomes of
the subgroup analysis for the mediated relationship model.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 731
subordinates’ trust in their superiors. The extant litera-
ture shows con?icting results when investigating the
relationship between formal control systems and trust
for collaborative team arrangements in general, and for
the relationships between superior and subordinate in
particular. We focused on superiors’ formal use of the
performance evaluation system, which we argued to be
an important underlying, generic aspect of the use of
such systems, and set out to explore the relationship be-
formality trust
feedback
quality
procedural
justice
Fig. 3. Results of the mediation model (front-of?ce subgroup, n = 75,
signi?cant paths displayed).
Table 7
Direct and indirect effects.
Path Direct relationship Partially mediated relationship Fully mediated relationship Q
2
R
2
Panel A: full sample, n = 160
a
Effect on TRUST 0.160 0.358
FORM > TRUST 0.247
***
0.208
**
0.148
PROCJ > TRUST 0.333
***
0.209
**
FEED > TRUST 0.412
***
Effect on FEED 0.044 0.080
FORM > FEED 0.254
**
Effect on PROCJ 0.127 0.233
FORM > PROCJ 0.130 0.066
FEED > PROCJ 0.413
***
Panel B: front-of?ce subsample (n = 75)
Effect on TRUST 0.162 0.325
FORM > TRUST 0.026 0.026 À0.006
PROCJ > TRUST 0.126 À0.035
FEED > TRUST 0.490
***
Effect on FEED 0.013 0.018
FORM > FEED 0.052
Effect on PROCJ 0.103 0.164
FORM > PROCJ 0.036 0.015
FEED > PROCJ 0.272
**
Panel C: back-of?ce subsample (n = 80)
Effect on TRUST 0.159 0.432
FORM > TRUST 0.370
***
0.284
**
0.208
PROCJ > TRUST 0.446
***
0.314
**
FEED > TRUST 0.343
***
Effect on FEED 0.071 0.105
FORM > FEED 0.314
**
Effect on PROCJ 0.141 0.213
FORM > PROCJ 0.222 0.103
FEED > PROCJ 0.397
***
**
p < 0.01 (based on t
(1000)
, two-tailed test).
***
p < 0.001 (based on t
(1000)
, two-tailed test).
a
The number of observations in the full sample (160) exceeds the sum of two subsamples (155) as ?ve managers have not reported their function and are
excluded from the subsample analysis.
formality
trust
feedback
quality
procedural
justice
Fig. 2. Results of the mediation model (full sample, n = 160, signi?cant
paths displayed).
formality trust
feedback
quality
procedural
justice
Fig. 4. Results of the mediation model (back-of?ce subgroup, n = 80,
signi?cant paths displayed).
732 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
tween formality of performance evaluation and interper-
sonal trust. Our ?ndings overall suggest that formality of
performance evaluation system use is indeed a relevant
characteristic in evoking subordinate trust. This ?nding
itself may explain why previous studies have found trust
to be enhanced by both ?nancial and non-?nancial per-
formance criteria that are both re?ections of formality
(Lau & Buckland, 2001; Lau & Shohilin, 2005), as both
suggest a formal use of metrics, rather than informal
evaluations (cf. Moers, 2005). For the overall model the
results con?rm the hypotheses that the link between for-
mality of the performance management cycle and trust is
mediated through the quality of feedback and procedural
justice (see Fig. 2). In particular, we ?nd that formality
matters for trust because it enhances the perceived
quality of feedback for subordinates. We found that high-
er perceived quality of feedback results in higher trust,
both directly, and via higher perceptions of procedural
justice.
In addition, we ?nd that the relationship between for-
mality of performance evaluation and trust is different
for managers in different situations, as attribution theory
predicts. We use the term contractibility to denote
differences in functions concerning easiness of de?ning
measurable outputs, and proxied for this variable by dis-
tinguishing front-of?ce managers, whose main responsi-
bilities regard market transactions, from back-of?ce
managers, whose main responsibilities regard internal
services. Managers in functions with highly contractible
outputs seem to be less affected by the level of formality
that their superiors apply in the performance evaluation
process than managers in functions with less measurable
outputs. These managers will see fair and transparent
performance feedback less as an outcome of their supe-
riors’ formal use of the system, but rather as a conse-
quence of the nature of their function and the
performance evaluation system (cf. Herold & Parsons,
1985). As their functional output itself provides formal
feedback immediately, little is attributable to the supe-
rior’s use of the PE system since superiors need little ef-
fort to formally use the available performance
information in their evaluation process (Sitkin & George,
2005). Managers in back-of?ce function, however, seem
affected by the formality that superiors apply in the per-
formance evaluation process. This means that the effort
superiors put in operationalizing output which by itself
is not easily measurable and traceable, results in trust
attributions.
For back-of?ce managers instead, formality has a po-
sitive effect on trust through both its effect on feedback
and justice. Feedback quality, however, appears to be
the most important factor in terms of the mediation ef-
fects explored. This implies that the link between perfor-
mance evaluation and trust is more complex than is
currently recognized in the procedural justice literature
that claims a central role for justice (e.g., Cohen-Charash
& Spector, 2003, p. 283). Indeed, when outputs are less
measurable, managers seem to perceive formal evalua-
tions, in which a superior puts less emphasis on his or
her subjective judgments, as more just. Our analysis
shows, however, that this is an indirect effect, which is
caused by the impact that improved performance
feedback has on justice perceptions. This is an important
?nding that suggest that studies on the relationship be-
tween performance evaluation systems and trust should
refocus on the feedback role of these systems (e.g., Lau
& Buckland, 2001; Lau & Shohilin, 2005). A ?nal differ-
ence between the two subgroups relates to the rele-
vance of justice for trust. For the front-of?ce managers,
justice and trust were unrelated, suggesting that other
superior’s behaviors, such as providing high-quality
feedback outside the performance evaluation cycle,
drives trust. We did not ?nd a signi?cant negative effect
of performance evaluation system use on trust as some
earlier studies reported. In these studies these negative
effects of controls on trust were found when controls
were newly installed (e.g., Tenbrunsel & Messick,
1999), and when existing formal and informal controls
were abolished (Malhotra & Munighan, 2002). Our study
instead focused on the continuous use of an existing set
of controls (i.e., the performance evaluation system),
applying a cross-sectional rather than longitudinal
analysis.
Finding differences between managerial groups is
important, since it not only suggests that the antecedents
of trust may be different across organizational functions,
but also explains previous contradictory results. First,
our ?nding of differences in trust attributions between
our two subgroups provides some reconciliation to the
opposite ?ndings of Coletti et al. (2005) and Malhotra
and Munighan (2002). Our study shows how context
may affect these attributions, which means that the nat-
ure or tasks may explain differential attributions in their
experimental settings. Note that both, in the earlier ?eld
studies (e.g., Lau & Buckland, 2001; Lau & Shohilin,
2005) and in the experimental studies (Coletti et al.,
2005; Tenbrunsel & Messick, 1999) research designs have
Table 8
Signi?cance of path differences.
a
Mediation model Front-of?ce Back-of?ce Difference T-test
Effect on TRUST
FORM > TRUST À0.006 0.208 À0.213 n.s.
PROCJ > TRUST À0.035 0.314
**
À0.349 À2.133
*
FEED > TRUST 0.490
***
0.343
***
0.147 n.s.
Effect on FEED
FORM > FEED 0.052 0.314
**
À0.262 À1.660
*
Effect on PROCJ
FORM > PROCJ 0.015 0.103 À0.088 n.s.
FEED > PROCJ 0.272
**
0.397
***
À0.125 n.s.
n.s., not signi?cant, df = 153.
*
p < 0.05 (two-tailed test).
**
p < 0.01 (two-tailed test).
***
p < 0.001 (two-tailed test).
a
Corresponding path coef?cients have been compared across back-
and front-of?ce using the following test t ¼
Path
fo
ÀPath
bo
???????????????????????????????????????????
ðmÀ1Þ
2
ðmþnÀ2Þ
ÂS:E
2
fo
þ
ðnÀ1Þ
2
ðmþnÀ2Þ
ÂS:E
2
bo
q h i
Â
???????
1
m
þ
1
n
p  Ã
where m is the sample size of front-of?ce,
n is the sample size of back-of?ce, fo is the front-of?ce, bo is the back-
of?ce, with m + n À 2 degrees of freedom.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 733
not explicitly taken into account the contractibility of
managerial outputs that we addressed through our sub-
group analyses.
9
This study is one of the few studies that investigate
the role of formal control systems in creating trust,
and one of the ?rst studies looking at hierarchical col-
laborative setting of the superior and the subordinate.
Obviously, the study is vulnerable to the typical weak-
nesses of survey research, regarding validity and reliabil-
ity of items and tests, although in the design and
execution we have avoided many of the potential pit-
falls. Both in the development and pre-test of the instru-
ment, and in the approach of respondents we followed
best practice. For some of the variables that we devel-
oped for this study, future studies may establish their
wider usefulness and validity. One ?nal potential limita-
tion of this study lies in our sample, in which we cannot
exclude the possibility that several subordinates have re-
ferred to the same superior in their answers. Although
this may have reduced the variance in scores, and there-
fore may have caused some bias, we believe reduced
variance in fact has decreased rather than increased
our chances of ?nding non-existent effects. Therefore,
we believe that our results cannot be explained by this
effect.
Future studies may look into the relationships between
performance evaluation systems and trust in further de-
tail. Clearly, it would be fruitful to replicate our survey
study using an experimental design, which is the method
used in much of the prior research. Such a design would
allow manipulating the level of contractibility across
experimental conditions, and con?rm the validity of our
proxy. It would also add a formal test of causality. Repli-
cation studies in other industries and functions would al-
low the asses the generalizability of our results. The study
provides numerous directions for future research. Future
studies could extend our study by analyzing effects of
other combinations of performance measurement
arrangements, and external conditions on trust. Overall,
however, we feel that the apparent complexity and rele-
vance of trust in the superior–subordinate setting pro-
vides a fruitful avenue for studies in performance
evaluation and control.
Acknowledgements
We thank participants at the annual EAA-conference,
the semi-annual AAA management accounting conference,
Kristy Towry, Ken Merchant, Henri Dekker, David Emsley,
Mike Shields, Simon C
?
adez? and seminar participants at
RSM Erasmus University, the University of
Ljubljana, and the University of Örebro for their comments
and suggestions towards earlier versions of this paper.
Appendix
Measurement instrument
Formality of performance evaluation system (FORM)
Question 1: Weighting of performance dimensions
This question is about the responsibilities of your job as
manager of a unit (e.g., a branch, or a department).
Below you ?nd four ways in which your supervisor may
assess your performance as unit manager. We are
interested in the relative importance of each of these
four performance areas. Therefore we ask you to allocate
a total number of 100 points to one or more of these
areas, re?ecting their relative importance. For example,
if you allocate the full 100 points to one area, you
express that only that area is important, and not the
other 3. If you allocate 25 points to each area, you
indicate that all areas are equally important for your
work, and that your superior does not have any priorities
or emphases. Please be as speci?c as possible in
allocating the 100 points, and answer these questions for
your function as unit manager, rather than for the bank
in general. There are no wrong or right answers. Note
that the areas are described broadly
Financial performance of your unit (e.g., sales,
pro?tability, interest margins, costs, value of
collected deposits, value of granted loans, value
of payment commissions, etc.)
Operational performance of your unit (e.g.,
productivity, completion of projects and phases
of projects, number of activities such as issued
credit cards, number of stipulated
contracts, etc.)
Employee relations within your unit (e.g., employee
satisfaction, turnover, achieved workforce
capabilities, number of trainings, etc.)
Quality and innovation of your unit (e.g., quality
awards, improvement of service quality,
customer satisfaction, introduction
of new services, ef?ciency
improvements, etc.)
Total (points or %) 100%
Question 2: Formality of target setting (FormTS)
Level of agreement with the following statements:
Less formal Scale More formal
When establishing
work objectives, my
superior does not
explicitly document
these objectives in
writing
1 2 3 4 5 When establishing
work objectives, my
superior explicitly
documents these
objectives in writing
When establishing
work objectives, my
superior casts these
objectives in
qualitative terms
1 2 3 4 5 When establishing
work objectives, my
superior casts these
objectives in
quantitative targets
9
The tasks that experimental subjects performed in the experimental
studies mentioned (Coletti et al., 2005; Malhotra & Munighan, 2002;
Tenbrunsel & Messick, 1999) differed widely in terms of complexity and
realism. As task design was not a treatment in any of these experimental
studies, explaining differences in trust attributions through task design
would be mere speculation.
734 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
This question was asked for all areas of performance, and
scores were averaged using the weights from question 1.
Question 3: Formality of performance measurement
(FormPM)
Level of agreement with the following statements:
Less formal Scale More formal
When judging my
performance, my
superior uses his
(her) personal
judgment of my
performance
1 2 3 4 5 When judging my
performance, my
superior relies on
objective information
from the information
system
When judging my
performance, my
superior discusses
my performance in
qualitative terms
1 2 3 4 5 When judging my
performance, my
superior expresses my
performance in
quantitative terms
(rates my performance
on quantitative rating)
This question was asked for all areas of performance, and
scores were averaged using the weights from question 1.
Question 4: Formality of rewarding (FormRew)
Level of agreement with the following statements:
Less formal Scale More formal
My ?xed pay is based on
my superior’s
personal judgment of
my performance
1 2 3 4 5 Objective
information from the
information system
My ?xed pay is based on
my performance in
qualitative terms
1 2 3 4 5 My performance in
quantitative terms
My bonus is based on
my superior’s
personal judgment of
my performance
1 2 3 4 5 Objective
information from the
information system
My bonus is based on
my performance in
qualitative terms
1 2 3 4 5 My performance in
quantitative terms
Question 5: Feedback quality (FEED)
Level of agreement with the statements (1 = I com-
pletely disagree, 2 = I disagree, 3 = neutral, 4 = I agree and
5 = I completely agree):
Feed
1
My superior gives me useful feedback about my job
performance
Feed
2
The performance information I get from my
superior is generally not very meaningful (reverse)
Feed
3
The feedback I receive from my superior helps me
do my job
Feed
4
I value the feedback I receive from my superior
Question 6: Procedural justice (PROCJ)
Level of agreement with the statements (1 = I com-
pletely disagree, 2 = I disagree, 3 = neutral, 4 = I agree and
5 = I completely agree):
Procj
1
I have full con?dence in the system’s fairness in
determining pay
Procj
2
I have full con?dence in the system’s fairness in
determining targets
Procj
3
I think that the way in which my performance is
measured is fair
Procj
4
I have full con?dence in the system’s fairness in
evaluating performance
Question 7: Trust in supervisor (TRUST)
Level of agreement with the statements (1 = I com-
pletely disagree, 2 = I disagree, 3 = neutral, 4 = I agree and
5 = I completely agree):
Trust
1
My superior will always act in my favor if he has
the chance
Trust
2
I am convinced that my superior will always fully
and honestly keep me up to date of everything
that is important to me
Trust
3
If my superior takes a decision that is against my
interest, I am convinced that this decision is
justi?ed for other reasons
Question 8: Task Uncertainty (UNCERT)
Level of agreement with the statements (1 = I com-
pletely disagree, 2 = I disagree, 3 = neutral, 4 = I agree and
5 = I completely agree):
Uncert
1
There is a clearly known way to do the major
types of work I normally encounter (reverse)
Uncert
2
There is a clearly de?ned body of knowledge of
subject matter which can guide me in doing my
work (reverse)
Uncert
3
There is an understandable sequence of steps
that can be followed in doing my work (reverse)
References
Bagozzi, R. P. (1994). Structural equation models in marketing research:
Basic principles. In Bagozzi (Ed.), Principles of marketing research.
Blackwell Business.
Birnberg, J., Frieze, I., & Shields, M. D. (1977). The role of attribution theory
in control systems. Accounting, Organizations and Society, 3, 189–200.
Blair, E., & Zinkhan, G. (2006). Nonresponse and generalizability in
academic research. Academy of Marketing Science, 34(1), 4–7.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable
distinction in social psychological research: Conceptual, strategic and
statistical considerations. Journal of Personality and Social Psychology,
51, 1173–1182.
Cahuc, P., & Zylberberg, A. (2004). Labour economics. Cambridge, MA: The
MIT Press.
Chenhall, R. H., & Lang?eld-Smith, K. (2003). Performance measurement
and reward systems, trust, and strategic change. Journal of
Management Accounting Research, 15, 117–143.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 735
Chin, W. W. (1998). The partial least squares approach to structural
equation modeling. In G. A. Marcoulides (Ed.), Modern methods for
business research (pp. 295–336). Mahwah, NJ: Lawrence Erlbaum
Associates, Publisher.
Clark, M. C., & Payne, R. L. (1997). The nature and structure of workers’
trust in managers. Journal of Organizational Behavior, 18, 205–224.
Cohen-Charash, Y., & Spector, P. E. (2003). The role of justice in
organizations: A meta-analysis. Organizational Behavior and Human
Decision Processes, 86, 278–321.
Coletti, A. L., Sedatole, K. L., & Towry, K. L. (2005). The effect of control
systems on trust in collaborative environments. The Accounting
Review, 80, 477–500.
Colquitt, J. A., & Jackson, J. L. (2006). Justice in teams: The context
sensitivity of justice rules across individual and team contexts. Journal
of Applied Social Psychology, 36, 868–899.
Cropenzano, R., Byrne, Z. S., Bobocel, D. R., & Rupp, D. E. (2001). Moral
virtues, fairness heuristics, social entities and other denizens of
organizational justice. Journal of Vocational Behavior, 58, 164–209.
Das, T. K., & Teng, B. S. (1998). Between trust and control: Developing
con?dence in partner cooperation in alliances. Academy of
Management Review, 23, 491–512.
Dekker, H. (2004). Control of inter-organizational relationships: Evidence
on appropriation concerns and coordination requirements.
Accounting, Organizations, and Society, 29, 27–49.
Dillman, D. A. (2000). Mail and Internet surveys: The tailored design method
(second ed.). New York, NY: John Wiley & Sons.
Dirks, K. T., & Ferrin, D. L. (2002). Trust in leadership: Meta-analytic
?ndings and implications for research and practice. Journal of Applied
Psychology, 87, 611–628.
Erdogan, B. (2002). Antecedents and consequences of justice perceptions
in performance appraisals. Human Resource Management Review, 12,
555–578.
Fine, G., & Holy?eld, L. (1996). Secrecy, trust, and dangerous leisure:
Generating group cohesion in voluntary organizations. Social
Psychology Quarterly, 59, 22–38.
Fisher, J. G., Maines, L. A., Peffer, S. A., & Sprinkle, G. B. (2005). An
experimental investigation of employer discretion in employee
performance evaluation and compensation. The Accounting Review,
80, 563–583.
Fornell, C., & Larcker, D. (1981). Evaluating structural equation models
with unobservable variables and measurement error. Journal of
Marketing Research, 18, 39–50.
Free, C. (2008). Walking the talk? Supply chain accounting and trust
among UK supermarkets and suppliers. Accounting, Organizations and
Society, 33, 629–662.
Gambetta, D. (Ed.). (1988). Trust: Making and breaking cooperative
relations. Oxford: Basil Blackwell.
Geisser, S. (1974). A predictive approach to the random effect model.
Biometrika, 61, 101–107.
Gibbs, M., Merchant, K. A., Van der Stede, W. A., & Vargus, M. E. (2004).
Determinants and effects of subjectivity in incentives. The Accounting
Review, 79, 409–436.
Greenberg, J. (1986). Determinants of perceived fairness of performance
evaluations. Journal of Applied Psychology, 71, 340–342.
Grossman, S. J., & Hart, O. D. (1986). The costs and bene?ts of ownership:
A theory of vertical and lateral integration. Journal of Political
Economy, 94, 691–719.
Hair, J., Anderson, R. E., Tatham, R. L., & Black, W. C. (1995). Multivariate
data analysis (fourth ed.). Englewood Cliffs, NJ: Prentice-Hall.
Hardin, R. (1992). The street-level epistemology of trust. Analyse & Kritik,
Zeitschrift fuer Sozialtheorie, 14, 152–176.
Hartmann, F. G. H. (2000). The appropriateness of RAPM: Toward the
further development of theory. Accounting, Organizations and Society,
25, 451–482.
Hartmann, F. G. H. (2005). The effects of tolerance for ambiguity and
uncertainty on the appropriateness of accounting performance
measures. ABACUS, 41, 241–264.
Hartmann, F. G. H. (2007). Do accounting performance measures indeed
reduce managerial ambiguity under uncertainty? Advances in
Management Accounting Research, 16, 159–180.
Herold, D. M., & Parsons, C. K. (1985). Assessing the feedback environment
in work organizations: Development of the job feedback survey.
Journal of Applied Psychology, 70, 290–305.
Hopwood, A. G. (1972). An empirical study of the role of accounting data
in performance evaluation. Journal of Accounting Research, 10,
156–182.
Ittner, C. D., & Larcker, D. F. (2001). Assessing empirical research in
managerial accounting: A value-based management perspective.
Journal of Accounting and Economics, 32, 349–410.
Ittner, C. D., Larcker, D. F., & Meyer, M. (2003). Subjectivity and the
weighting of performance measures: Evidence from a balanced
scorecard. The Accounting Review, 78(3), 725–758.
Jones, T. M. (1995). Instrumental stakeholder theory: A synthesis of ethics
and economics. Academy of Management Review, 20, 404–437.
Kramer, R. M. (1999). Trust and distrust in organizations: Emerging
perspectives, enduring questions. Annual Review of Psychology, 50,
569–598.
Landy, F. J., Barnes, J. L., & Murphy, K. R. (1978). Correlates of perceived
fairness and accuracy of performance evaluation. Journal of Applied
Psychology, 63, 751–754.
Lau, C. M., & Buckland, C. (2001). Budgeting – Role of trust and
participation: A research note. ABACUS, 37, 369–388.
Lau, C. M., & Shohilin, M. (2005). Financial and non?nancial performance
measures: How do they affect job satisfaction. British Accounting
Review, 37, 389–413.
Lau, C. M., & Tan, S. L. C. (2006). The effects of procedural fairness and
interpersonal trust on job tension in budgeting. Management
Accounting Research, 17, 171–186.
Leventhal, G. S. (1976). Fairness in social relationships. In J. W. Thibaut, J.
T. Spence, & R. C. Carson (Eds.), Contemporary topics in social
psychology (pp. 211–239). Morristown, New York: General Learning
Press.
Leventhal, G. S., Karuza, J., & Fry, W. R. (1980). Beyond fairness: A theory
of allocation preferences. In G. Mikula (Ed.), Justice and social
interaction (pp. 167–218). New York: Springer-Verlag.
Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task
performance. Englewood-Cliffs: Prentice-Hall.
Lohmöller, J. B. (1989). Latent variable path modeling with partial least
squares. Heidelberg: Physica-Verlag.
Luft, J., & Shields, M. D. (2003). Mapping management accounting:
Graphics and guidelines for theory-consistent empirical research.
Accounting, Organizations and Society, 28, 169–249.
Malhotra, D., & Munighan, J. K. (2002). The effects of contracts on
interpersonal trust. Administrative Science Quarterly, 47, 534–559.
March, J. G., & Olsen, J. P. (1989). Rediscovering institutions. New York: Free
Press.
Marginson, D., & Ogden, S. (2005). Coping with ambiguity through the
budget: The positive effects of budgetary targets on managers’
budgeting behaviours. Accounting, Organizations and Society, 30,
435–456.
Moers, F. (2005). Discretion and bias in performance evaluation: The
impact of diversity and subjectivity. Accounting, Organizations and
Society, 30, 67–80.
Morand, D. D. (1995). The role of behavioral formality and informality in
the enactment of bureaucratic versus organic organizations. Academy
of Management Review, 20, 831–872.
Otley, D. T. (1999). Performance management: A framework for
management control systems research. Management Accounting
Research, 10, 363–382.
Read, W. H. (1962). Upward communication in industrial hierarchies.
Human Relations, 15, 3–15.
Ross, A. (1994). Trust as a moderator of the effect of performance
evaluation style on job-related tension: A research note. Accounting,
Organizations and Society, 19, 629–635.
Rousseau, D. M., Sitkin, S. B., Byrt, R. S., & Camerer, C. (1998). Not so
different after all: A cross-discipline view of trust. Academy of
Management Review, 23, 393–404.
Sitkin, S. B., & George, E. (2005). Managerial trust-building through the
use of legitimating formal and informal control systems. International
Sociology, 20, 307–338.
Steelman, L. A., Levy, P. E., & Snell, A. F. (2004). The feedback environment
scale: Construct de?nition, measurement and validation. Educational
and Psychological Measurement, 64, 165–184.
Stone, M. (1974). Cross validity choice and assessment of statistical
predictions. Journal of the Royal Statistical Society, Series B, 36,
111–133.
Taylor, M. S., Tracy, K. B., Renard, M. K., Harrison, J. K., & Carroll, S. J.
(1995). Due process in performance appraisal: A quasi-experiment in
procedural justice. Administrative Science Quarterly, 40, 495–
523.
Taylor, M. S., Masterson, S. S., Renard, M. K., & Tracy, K. B. (1998).
Managers’ reactions to procedurally just performance management
systems. Academy of Management Journal, 41, 568–579.
Tenbrunsel, A. E., & Messick, D. M. (1999). Sanctioning systems, decision
frameworks, and cooperation. Administrative Science Quarterly, 44,
684–707.
Tenenhaus, M., Vinzi, V. E., Chatelin, Y., & Lauro, C. (2005). PLS path
modeling. Computational Statistics & Data Analysis, 48, 159–205.
736 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
Tomkins, C. (2001). Interdependencies, trust and information in
relationships, alliances and networks. Accounting, Organizations and
Society, 26, 161–191.
Vandenbosch, M. B. (1996). Con?rmatory compositional approaches to
the development of product spaces. European Journal of Marketing, 30,
23–46.
Whitener, E. M., Brodt, S. E., Korsgaard, M. A., & Werner, J. M. (1998).
Managers as initiators of trust: An exchange relationship framework
for understanding managerial trustworthy behavior. Academy of
Management Review, 23, 513–530.
Withey, M., Daft, R. L., & William, H. C. (1983). Measures of Perrow’s work
unit technology: An empirical assessment and a new scale. The
Academy of Management Journal, 26(1), 45–63.
Williamson, O. E. (1993). Calculativeness, trust, and economic
organization. The Journal of Law and Economics, 36, 453–486.
Wold, H. (1982). Soft modeling – The basic design and some extensions. In
K. Jöreskog & H. Wold (Eds.), Systems under indirect observation II
(pp. 1–53). Amsterdam: North-Holland Press.
Wold, H. (1985). Partial least squares. In S. Kotz & N. L. Johnson (Eds.).
Encyclopedia of statistical sciences (vol. 6, pp. 581–591). New York:
Wiley.
Zaheer, A., McEvily, B., & Perrone, V. (1998). Does trust matter? Exploring
the effects of interorganizational and interpersonal trust on
performance. Organization Science, 9, 141–159.
Zimmerman, J. L. (2003). Accounting for decision making and control.
Boston: Irwin, McGraw-Hill.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 737
doc_323975668.pdf
An important question in the control literature concerns the role of interpersonal trust in
the design and functioning of formal control systems for collaborative settings. In this
paper, we argue that subordinate’s trust in the superior depends on the formality of the
performance evaluation procedure and that this relationship is mediated by managerial
perceptions of justice and feedback. The findings confirm our predictions. Furthermore,
we find that formality matters more for trust formation for those managers that are in
functions with less contractible outputs.
How formal performance evaluation affects trust between superior
and subordinate managers
Frank Hartmann
a,1
, Sergeja Slapnic?ar
b,
*
a
RSM Erasmus University, Rotterdam Burgemeester Oudlaan 50, 3062 PA Rotterdam, The Netherlands
b
University of Ljubljana, Faculty of Economics, Kardeljeva pl. 17, 1000 Ljubljana, Slovenia
a b s t r a c t
An important question in the control literature concerns the role of interpersonal trust in
the design and functioning of formal control systems for collaborative settings. In this
paper, we argue that subordinate’s trust in the superior depends on the formality of the
performance evaluation procedure and that this relationship is mediated by managerial
perceptions of justice and feedback. The ?ndings con?rm our predictions. Furthermore,
we ?nd that formality matters more for trust formation for those managers that are in
functions with less contractible outputs.
Ó 2008 Elsevier Ltd. All rights reserved.
Introduction
In contemporary business environments, interpersonal
trust is considered an important contributor to organiza-
tional performance (e.g., Dirks & Ferrin, 2002; Zaheer,
McEvily, & Perrone, 1998). Interpersonal trust is important
whenever organizational participants engage in situations
that demand collaborative effort (cf. Coletti, Sedatole, &
Towry, 2005). It is especially important in supervisor–sub-
ordinate relationships, as trust improves cooperation and
mitigates agency problems (e.g., Jones, 1995) by enhancing
information exchange and reducing subordinates’ per-
ceived need to engage in short term opportunistic behavior
(e.g., Fisher, Maines, Peffer, & Sprinkle, 2005). Although
important, interpersonal trust is generally not deemed
suf?cient for supervisors to control their subordinates’
behaviors. Therefore, supervisors typically use formal
controls such as performance evaluation systems (cf. Das
& Teng, 1998; Malhotra & Munighan, 2002). In this paper,
we investigate whether and how superiors’ use of perfor-
mance evaluation systems affects subordinates’ trust in
their superiors.
This investigation is warranted since the relationship
between performance evaluation and interpersonal trust
is not well understood. For example, Lau and Buckland
(2001) and Lau and Shohilin (2005) both investigated the
trust effects of superiors’ choice of ?nancial or non-?nan-
cial performance metrics for evaluating subordinate per-
formance, but report almost opposite ?ndings. Whereas
the former study concludes that trust is higher when supe-
riors use ?nancial performance measures, which are more
‘objective’ and ‘truthful’, the latter study reports that trust
is higher when superiors use non-?nancial measures,
which are ‘broader’ and ‘more complete’. Research in team
settings also provides mixed evidence on the effects of for-
mal controls on trust. Studies on trust attributions (e.g.,
Malhotra & Munighan, 2002; Tenbrunsel & Messick,
1999) suggest that formal control systems may decrease
interpersonal trust as they cause collaborative behavior
to be attributed to the presence of formal controls, rather
than to team member trustworthiness. In contrast, Coletti
et al. (2005) ?nd positive trust effects of using formal per-
formance evaluation systems in teams. They argue that
these systems are essential in providing collaborators with
performance feedback, which contain implicit signals
about mutual trustworthiness.
In combination, available evidence suggests that perfor-
mance evaluation systems have a potentially important
impact on trust, but also raises questions about which
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.11.004
* Corresponding author. Tel.: +386 1 5892 483; fax: +386 1 5892 698.
E-mail addresses: [email protected] (F. Hartmann), Sergeja.slapni-
[email protected] (S. Slapnic?ar).
1
Tel.: +31 (0) 10 40 82 222; fax: +31 (0) 10 40 89 017.
Accounting, Organizations and Society 34 (2009) 722–737
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
aspects of performance evaluation systems affect trust and
how, and about how the attribution effects that such sys-
tems have may affect the creation of trust.
In this paper, we address these questions by developing
and testing a path model that relates superior’s use of the
performance evaluation system to subordinate’s subse-
quent trust in the superior. To reconcile opposite ?ndings
from studies that focused on superiors’ choice of metrics
(e.g., Lau & Buckland, 2001; Lau & Shohilin, 2005), we pro-
pose that trust is associated with superiors’ formal use of
the performance evaluation system. We then explore two
paths by which superior’s formal use of the performance
evaluation system may affect subordinate’s trust: (i)
through the enhanced perceived justice of the performance
evaluation process and (ii) through the enhanced perceived
quality of performance feedback. We ?nally investigate
whether the path-relationships are stable across manage-
rial functions, arguing that the attribution of trustworthy
behavior to the superior, rather than the performance eval-
uation system, depends on managerial job characteristics.
By developing and testing this model our contribution
to the literature is fourfold. First, we aim to show that trust
is affected by the degree of formality in the use of perfor-
mance measures rather than the type of measures, thus
reconciling the opposite results from studies that looked
at the choice of performance metrics (Lau & Buckland,
2001; Lau & Shohilin, 2005). Second, by exploring two
mediators of the relationship between performance evalu-
ation system use and trust, we provide an explicit test of
the intervening mechanisms that previous studies merely
suggested but never addressed empirically. Third, we ex-
plain differences in trust attributions by differences in
managerial job characteristics in an attempt to reconcile
related ?ndings in previous studies (Coletti et al., 2005;
Malhotra & Munighan, 2002). Fourth and ?nally, this paper
integrates the literature on control and trust attributions in
team settings (e.g., Coletti et al., 2005) with the literature
on the hierarchical setting of superiors and subordinates,
the latter of which is an area of primary concern to the
management accounting literature at large (Hartmann,
2000, 2007).
The remainder of this paper is structured as follows.
Section ‘‘Literature review” provides an overview of the ex-
tant literature on the relationships between trust and per-
formance evaluation systems and identi?es some of the
gaps and con?icts in extant studies. Section ‘‘Hypotheses
development” develops the path model from which
hypotheses are derived about the relationship between
supervisory use of the performance evaluation system
and trust. Section ‘‘Method” presents the design of the
empirical survey that we conducted. Section ‘‘Results” pre-
sents the results from the tests of the hypotheses. Section
‘‘Discussion and conclusions” concludes this paper with
an overview of ?ndings, the study’s limitations and some
directions for future research.
Literature review
Despite the large variety of de?nitions and descriptions
of trust in organizational settings, both the organizational
and accounting literatures generally recognize the impor-
tance of trust for organizational functioning (cf. Coletti
et al., 2005; Rousseau, Sitkin, Byrt, & Camerer, 1998). In
accounting studies, trust appears to be an important factor
across organizational settings and levels of analysis, and
explains the functioning of company-wide pay-for-perfor-
mance systems as well as the success of inter-organiza-
tional networks (e.g., Chenhall & Lang?eld-Smith, 2003;
Dekker, 2004; Free, 2008; Tomkins, 2001). In this study,
we focus on the interpersonal trust of subordinates’ in
their superiors, which has received some attention in the
accounting literature over the last decades (e.g., Hopwood,
1972; Lau & Tan, 2006). We de?ne trust in this setting as
the belief that subordinates hold about the trustworthiness
of their superiors (cf. Clark & Payne, 1997; Coletti et al.,
2005). In accordance with the psychological view on trust,
we consider a person’s trustworthiness as that person’s
preference for upholding social and behavioral norms, such
as honesty and fairness, in situations in which the trusting
party will bene?t from the adherence to those norms (cf.
Rousseau et al., 1998).
2
The few studies that investigated the effects of supe-
rior’s performance evaluation behavior on subordinate’s
trust thus far seem to con?rm that upholding social norms
of honesty and fairness enhances trust, but yield con?ict-
ing results concerning the role of the performance evalua-
tion system (e.g., Hopwood, 1972; Lau & Buckland, 2001;
Lau & Shohilin, 2005; Ross, 1994). These studies have
mainly focused on the superior’s choice of performance
metrics for evaluating and rewarding subordinate perfor-
mance. Hopwood (1972), for example, found that subordi-
nates reported higher trust in their superiors, when these
superiors measured subordinates’ performance strictly
against budget. This ?nding was later con?rmed by Lau
and Buckland (2001), who showed that trust mediates
the relationship between the use of budget-based perfor-
mance metrics and job-related tension. In both studies,
these ?ndings were explained by the fact that subordinates
perceive the use of budget-related performance criteria as
the superior’s attempt to provide ‘precise’ and ‘honest’ per-
formance feedback. In contrast, in these studies the use of
non-?nancial performance metrics was associated with
lower levels of trust. The authors argued that their use
was perceived as ‘subjective’ and ‘inaccurate’, which is sur-
prising as no a priori reason exist why ?nancial metrics
would be more subjective and accurate than other (non-
?nancial) quantitative metrics.
Although these results seem to generally support the
importance of upholding norms of honesty and clarity for
superiors’ trustworthiness (cf. Clark & Payne, 1997;
2
The economic literature considers trust as the outcome of a calculation-
based prediction a trusting party makes about a trusted party’s future
behavior (see, e.g., Hardin, 1992; Kramer, 1999; Williamson, 1993). The
psychological notion of trust adopted here considers trust as the outcome
of affective and emotional considerations that form the trusting party’s
belief about the trusted party’s behavioral dispositions (cf. Fine & Holy?eld,
1996; March & Olsen, 1989). The psychological view on trust disputes an
important role of calculation in trust creation, as it considers the associated
calculations too complex for the cognitive capacities of any trusting party
(cf. Fine & Holy?eld, 1996).
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 723
Rousseau et al., 1998), further evidence on the relationship
between performance evaluation system use and trust
raises doubt about the conclusiveness of these ?ndings.
First, the results of Hopwood (1972) and Lau and Buckland
(2001) seem contradicted by the study of Lau and Shohilin
(2005). Lau and Shohilin (2005) followed a similar line of
reasoning, but found in a different sample that trust was
higher when superiors used non-?nancial performance
metrics. Their explanation was that non-?nancial metrics
have a higher accuracy as they enable a ‘broader’ perfor-
mance assessment in the eyes of the subordinates than
the use of ‘narrow and rigid’ budget-based metrics (Lau &
Shohilin, 2005, p. 396). Thus, although it seems undisputed
that upholding social norms may evoke trust in evaluative
settings, these ?ndings suggest that perceived adherence
to such norms is not related to the use of a particular kind
of metrics, and that different aspects of performance eval-
uation system use may be involved.
Second, studies on the attribution effects of control sys-
tems show that upholding norms of honesty and clarity by
a superior does not necessarily result in perceptions of
higher trustworthiness (e.g., Malhotra & Munighan, 2002;
Tenbrunsel & Messick, 1999). Indeed, when honest and
clear behavior is perceived to be caused by an external
control system, rather than a person’s inherent intentions,
trust will not emerge as such behavior is not attributed to
trustworthiness (Gambetta, 1988; Malhotra & Munighan,
2002). This is a central claim of attribution theory, which
is generally concerned with the question as to whether a
person explains the behavior of another person by situa-
tional factors or by the other person’s inherent dispositions
(cf. Birnberg, Frieze, & Shields, 1977; Gambetta, 1988). This
theory predicts that upholding social norms will only re-
sult in trust attributions when the person makes a costly
choice by upholding such norms, as this signals that the
person’s upholding of norms is intentional (cf. Gambetta,
1988; Malhotra & Munighan, 2002, p. 537). In their study
on team collaboration with and without formal controls,
Malhotra and Munighan (2002) thus expected and found
that using binding (i.e., formal) team contracts to control
team cooperation led to lower intra-team trust than the
use of non-binding (i.e., informal) team contracts. Binding
contracts caused team members to attribute cooperation
to the formal contract, while a voluntary adherence to
the non-binding contract was seen as a signal of inherent
trustworthiness. In their study on performance evaluation
systems in a team setting, Coletti et al. (2005) demon-
strated that such negative attribution effects may be com-
pensated by positive attribution effects of performance
evaluation system use on trust. They found that interper-
sonal trust increased when the performance evaluation
system enhanced cooperative performance, and when the
system provided team members with feedback on their
joint performance.
Overall, empirical evidence on performance evaluation
and trust strongly suggests that they are related, yet fails
to provide conclusive evidence on the aspect of the perfor-
mance evaluation system that matters and on the attribu-
tions that these systems may cause. In section ‘‘Hypotheses
development”, we will explore these two issues by devel-
oping set of related empirical hypotheses about the rela-
tionship between the use of performance evaluation
systems and trust.
Hypotheses development
The opposite ?ndings of studies that investigated the ef-
fect of performance metrics on trust (e.g., Lau & Buckland,
2001; Lau & Shohilin, 2005) can be reconciled by acknowl-
edging that in these studies trust may have been the con-
sequence of the performance measurement process,
rather than the type of performance metric used. The per-
formance evaluation literature indeed con?rms that supe-
riors differ in the extent to which they use available
systems and choose to apply explicit or implicit measure-
ment (Hartmann, 2000; Moers, 2005). This choice may en-
hance trust, as it signals the superior’s intention to uphold
norms of honesty and clarity, which only explicit measures
can provide (cf. Lau & Buckland, 2001; Lau & Shohilin,
2005). However, since performance evaluation is broader
than performance measurement alone, subordinates’ trust
responses are likely to be a reaction to all stages of the
performance evaluation process (Dirks & Ferrin, 2002).
Thus, to explain trust reactions to performance evaluation
system use, our analysis will also consider the choices that
superiors make concerning the development of perfor-
mance targets and the way in which superiors assign
performance related rewards (cf. Fisher et al., 2005;
Greenberg, 1986; Ittner & Larcker, 2001; Otley, 1999;
Zimmerman, 2003).
In line with the de?nitions of formal and informal con-
trol in the organizational literature (see, e.g., Morand,
1995; Sitkin & George, 2005), we propose to address these
choices in terms of the level of formality that superiors ap-
ply when using the performance evaluation system.
3
For
the different stages of the performance evaluation process
formality has the following meaning. Regarding target set-
ting, a high level of formality denotes the situation in which
superiors explicate performance targets in quantitative and
written terms. Informal target setting, instead, means having
implicit, qualitative (‘do your best’) targets, that are commu-
nicated informally, and whose subsequent achievement can-
not be measured objectively (cf. Locke & Latham, 1990).
Regarding performance measurement, a high level of for-
mality implies the use of quantitative and objective, rather
than qualitative and subjective measures, which make up
informal performance assessments (cf. Moers, 2005).
4
3
In the wider organizational literature the terms formal control and
informal control have various connotations beyond our strict de?nition.
‘Informal control’ is seen as ‘loose’, ‘spontaneous’ and ‘casual’. ‘Formal
control’ instead re?ects as ‘tight’, ‘deliberate’ and ‘bureaucratic’ (see, e.g.,
Morand, 1995, p. 831).
4
Moers’ (2005, p. 71) provides little theoretical justi?cation for his
distinction between ‘objective’ and ‘subjective’ measures. Therefore, it is
not clear whether the distinction relates to the measure’s properties, to the
process by which they are derived, nor to whose accord the measure should
be seen as ‘objective’ or ‘subjective’. According to the examples given in the
paper, ‘objective measures’ seem to be those that include some explicit
numerical standard (e.g., ‘2% cost decrease’), whereas ‘subjective’ measures
simply refer to activities (e.g., good use of resources’). We believe that this
distinction corresponds with our more precise distinction between high
and low levels of formality.
724 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
Regarding the assignment of performance-based rewards, a
high level of formality means the formulaic determination
of rewards and bonuses, rather than the use of untraceable
personal judgment (cf. Gibbs, Merchant, Van der Stede, &
Vargus, 2004). In sum, superiors that use the performance
evaluation in a formal way are those superiors that explicate
performance targets, measure performance by a set of clear
metrics, and who give rewards based on clear allocation
rules. We regard these three phases of the performance eval-
uation process as three-dimensions of the performance eval-
uation system that determine its overall formality (see
Fig. 1).
We expect that a formal use of the performance evalu-
ation system, as de?ned above, will have a positive effect
on trust, as it allows higher levels of integrity, honesty,
accuracy and consistency in performance evaluation than
informal performance evaluation. This is not only consis-
tent with the studies by Hopwood (1972), Lau and
Buckland (2001) and Lau and Shohilin (2005), but also with
other recent evidence on the behavioral effects of implicit
and subjective performance measures (e.g., Hartmann,
2007; Moers, 2005). Moers (2005), for example, recently
shows that subjective performance evaluations result in
biases, such as ‘compression’ (i.e., low-rating variance),
and ‘leniency’ (i.e., high-rating average), which are clearly
at odds with the social norms of honesty, accuracy and
consistency that form the necessary conditions for subor-
dinates’ trust in their superior (e.g., Rousseau et al., 1998;
Whitener, Brodt, Korsgaard, & Werner, 1998). In their
study on the trust effects of managerial decision-making,
Sitkin and George (2005) show that when superiors have
to make painful or controversial personnel decisions they
tend to counter threats to their trust and legitimacy by
increasing their use of formal controls and reducing their
use of informal controls.
Before exploring more detailed relationships between
performance evaluation system use and trust, we therefore
propose to test the following relationship as an overall
starting point of our analysis:
H1: Formal use of the performance evaluation system
by superiors has a positive effect on subordinates’ trust
in superiors.
Formality, trust and the mediating effect of procedural justice
Trust is a consequence of managers’ adhering to social
norms as integrity, honesty, accuracy and consistency (cf.
Clark & Payne, 1997, p. 208; Whitener et al., 1998). We ex-
pect that these factors mediate the relationship between
performance evaluation and trust (H1) because they deter-
mine the procedural justice of the performance evaluation
process.
The organizational justice literature demonstrates that
the social norms that evoke trust show great resem-
blance to the social norms that determine whether orga-
nizational decision-making processes will be judged ‘fair’
by the stakeholders of these processes (e.g., Leventhal,
1976; Leventhal, Karuza, & Fry, 1980). Leventhal (1976,
p. 39), for example, argues that decision-makers should
adhere to such social norms as consistency, freedom
from bias, and accuracy to enhance the procedural
justice of the decision-making processes they engage in.
Therefore, when superiors uphold the social norms of
honesty, accuracy and consistency, procedural justice is
enhanced and trust is likely to emerge (e.g., Rousseau
et al., 1998; Whitener et al., 1998). In their meta-analytic
review Dirks and Ferrin (2002) con?rm the positive rela-
tionship between procedural justice and trust in extant
research.
Since we expect that procedural justice mediates the
relationship between performance evaluation and trust,
we propose that procedural justice is a direct and positive
consequence of superiors’ formal rather than informal use
of the performance evaluation system when making per-
formance evaluation decisions. Superiors who explicate
performance targets, measure performance by clear met-
rics, and give rewards based on clear allocation rules (i.e.,
high formality) will be able to deliver more consistent,
more accurate and less biased performance evaluations
(i.e., higher procedural justice) than superiors who use
implicit targets, measure performance subjectively and
allocate rewards using mere personal judgment. This is
consistent with the limited and partial evidence from the
accounting literature (e.g., Hopwood, 1972; Lau &
Buckland, 2001), as well as with ?ndings in some early
organizational studies that documented the fairness conse-
formality trust
FormTS
FormP
Formrew
Procj1 Procj2 Procj3
Feed1 Feed2 Feed3
Trust1
Trust2
Trust3
Procj3
Feed3
feedback
quality
procedural
justice
uncertainty
tenure
Fig. 1. Full model (all predicted relationships are positive).
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 725
quences of performance evaluation process characteristics
(e.g., Greenberg, 1986; Landy, Barnes, & Murphy, 1978).
5
Colquitt and Jackson (2006, p. 871) recently found that fair-
ness criteria should apply to all stages of a decision-making
process, such as the information search, the choice of deci-
sion rules, and the communication of outcomes. This sup-
ports our expectation that trust results from superiors
acting consistently, accurately and honestly throughout the
whole performance evaluation process, rather than giving
subordinates a favorable evaluation as such (Taylor, Master-
son, Renard, & Tracy, 1998; Taylor, Tracy, Renard, Harrison,
& Carroll, 1995).
Overall, therefore, we expect that when superiors expli-
cate performance targets in formal terms, attempt to quan-
tify measurement and allocate rewards using formulaic
and traceable procedures, this enhances procedural justice
and therefore trust. This is expressed in the following
hypothesis:
H2: The positive effect of the formal use of the perfor-
mance evaluation system on trust is mediated by per-
ceived procedural justice, such that the formal use of
the performance evaluation system has a positive effect
on perceived procedural justice, which in turn has a
positive effect on trust.
Formality, trust and the mediating effect of performance
feedback quality
Our review of accounting studies on trust antecedents
suggests a second mechanism by which performance eval-
uation procedures affect trust, which is via the perfor-
mance feedback that such evaluation systems typically
provide (e.g., Coletti et al., 2005; Lau & Buckland, 2001). In-
deed, these studies suggest that the effect of performance
evaluation on trust may be mediated by the increased
quality of performance feedback that subordinates obtain,
in cases when superiors evaluate performance formally,
rather than informally. Performance feedback is de?ned
as the feedback that subordinates get about the perfor-
mance levels they attain (cf. Steelman, Levy, & Snell,
2004). Such feedback is considered an essential function
of any performance evaluation process, as it should help
subordinates understand how they may adapt their behav-
iors to enhance performance (e.g., Herold & Parsons, 1985;
Steelman et al., 2004), and thus to improve collaboration
with their superior. Last, but not least, good feedback al-
lows subordinates to adjust the expectations about their
rewards (Cahuc & Zylberberg, 2004). Although many fac-
tors affect feedback quality (cf. Herold & Parsons, 1985),
the management literature strongly suggests that effective
feedback on individual performance requires that targets
are set clearly (e.g., Hartmann, 2007; Locke & Latham,
1990), that performance is measured accurately and in
an understandable way (Steelman et al., 2004), and that
reward distribution is transparent and traceable (cf. Moers,
2005). Formal evaluations have an advantage over infor-
mal evaluations in that they clearly specify the dimensions
of performance that are being evaluated, and the way in
which the rewards are linked to those performance dimen-
sions (cf. Marginson & Ogden, 2005; Moers, 2005), which
enhances goal orientation and motivation (cf. Hartmann,
2005; Hartmann, 2007). Thus, we predict that formal eval-
uations are positively related to trust because they provide
subordinates with a higher quality of feedback than do
informal evaluations (cf. Lau & Buckland, 2001; Moers,
2005). Coletti et al. (2005) show that receiving high-qual-
ity feedback will positively affect trust, as subordinates
will perceive superiors’ attempts to provide good feedback
as an indication of superiors’ desire to help subordinates
improve their performance as well as to improve their
working relations. These factors are all part of the social
norms underlying trustworthiness (Rousseau et al.,
1998). Consistent with extant evidence on the relation-
ships between performance evaluation, feedback and trust,
we therefore predict that the overall positive effect of for-
mality on trust is mediated by the quality of performance
feedback as well. This is summarized in the following
hypothesis:
H3: The positive effect of the formal use of the perfor-
mance evaluation system on trust is mediated by the
perceived quality of performance feedback, such that
the formal use of the performance evaluation system
has a positive effect on perceived quality of perfor-
mance feedback, which in turn has a positive effect on
trust.
It is clear that the theory underlying hypotheses 1, 2
and 3 suggests relatedness between feedback quality and
procedural justice, since the underlying characteristics
share several similarities. The factors such as accuracy
and lack of bias seem equally important for evoking per-
ceptions of procedural justice, as well as for providing
high-quality feedback. In large parts of the organizational
justice literature, this relatedness has been recognized,
such that performance feedback itself is considered an
important antecedent of procedural justice (e.g., Cohen-
Charash & Spector, 2003; Taylor et al., 1995). Thus,
according to these studies on procedural justice, good
performance feedback is a constitutive element of a fair
(evaluation) process, as it provides the openness and high
level of communication (e.g., Colquitt & Jackson, 2006;
Erdogan, 2002) that fair processes demand. We therefore
expect that perceived quality of performance feedback
not only has a direct effect on trust, as formulated in
hypothesis 2, but also affects trust via procedural justice.
We therefore propose to test the following hypothesis as
part of the overall model:
H4: The perceived quality of performance feedback
has a positive impact on perceived procedural
justice.
5
The survey study by Landy et al. (1978) reports positive correlations
between ‘fairness’ of performance evaluation and the existence of a formal
evaluation program, and formal goal setting. The survey study by Green-
berg (1986) reports ‘consistent application of standards’ as an important
correlate of procedural justice of performance appraisals.
726 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
Formality of performance evaluation, trust and managerial
job characteristics
Our expectations so far implicitly assume that when
subordinates receive fair and high-quality performance
evaluations, at least some of these qualities will be attrib-
uted to the superior providing the evaluations. This is con-
sistent with available empirical evidence (e.g., Coletti et al.,
2005; Hopwood, 1972; Lau & Buckland, 2001). However, to
explicitly account for attribution effects that formal control
systems may cause, we now explore whether the relation-
ships identi?ed are likely to be affected by such
attributions.
Generally, attribution theory suggest that the emer-
gence of trust requires that subordinates perceive superi-
ors’ provision of fair and high-quality performance
evaluations as a signal of their inherent trustworthiness,
rather than as a consequence of the formal control system
in place (cf. Malhotra & Munighan, 2002). In other words,
although subordinates may be generally inclined to trust
superiors who provide honest, accurate and consistent per-
formance evaluation and clear performance feedback, the
effect on trust will be higher if subordinates believe that
these behaviors are a consequence of superior’s deliberate
and costly intentions.
As regards the positive effects of the formal use of the
performance evaluation system, expressed in hypotheses
1–3, we expect that the extent to which the formal use
of the performance evaluation system results in trust, via
perceptions of higher fairness and higher quality of perfor-
mance feedback, may depend on managerial job character-
istics. Based on attribution theory, we expect that the
extent to which subordinates will attribute superior’s for-
mal use of the performance evaluation to their inherent
trustworthiness depends on the extent to which formal
performance evaluation is costly to the superior, that is,
when it requires high levels of effort of the superior (cf.
Gambetta, 1988; Herold & Parsons, 1985; Malhotra &
Munighan, 2002). We argue that such will be the case for
those subordinates’ functions for which outputs are not
easily measurable in quantitative and objective terms.
For brevity purposes, we use the term contractibility to
capture the ability to measure outputs (cf. Grossman &
Hart, 1986). Examples of high-contractibility outputs
include typical ?nancial and quantitative indicators of
turnover and pro?t, and other technical outputs for which
measures and targets are relatively easy to develop
(cf. Moers, 2005). Such outputs are widely available in
organizational functions where subordinates are engaged
in functions related to external trade and sales. Examples
of low-contractibility outputs include those in which
measurability of individual performance is hindered by
organizational interdependencies, and in which job tech-
nology causes a lack of objective metrics. This refers to
functions in which subordinates are responsible for provid-
ing internal service, advice and support (cf. Moers, 2005).
In jobs for which outputs are not easily contractible, supe-
riors have to apply extra effort to explicate performance
targets, measure performance clearly, and give rewards
based on clear allocation rules than in jobs with easily
contractible outputs (Hartmann, 2007). For our model this
implies the following.
Regarding procedural justice, we expect that subordi-
nates in functions with relatively contractible outputs
(e.g., in sales functions) will be less prone to attribute per-
formance evaluation formality to superiors’ inherent trust-
worthiness. As superiors can use easily available metrics
for target setting, performance measurement and reward
distribution for these functions, formality does not provide
a strong signal to subordinates about superiors’ true and
costly intentions toward providing fair evaluations. Thus,
formality will add less to fairness attributions in such func-
tions, than in functions in which formality requires more
effort on the part of the superior (cf. Sitkin & George,
2005). This situational dependency of fairness judgment
is consistent with recent ?ndings by Colquitt and Jackson
(2006), who showed that fairness judgment may indeed
depend on context.
Regarding the second mediating variable, performance
feedback, a similar argument applies. In situations where
subordinates already have good understanding of their
own performance levels and the consequences of this per-
formance level for individual wealth, the bene?ts of formal
performance evaluation in terms of providing feedback
information will be less prevalent. Here, attribution theory
suggests that the outcomes of formal ways of performance
evaluation are likely to be considered the consequence of
the formal system, rather than the outcome of supervisor’s
deliberate effort to provide such information (cf. Sitkin &
George, 2005). In contrast, subordinates in functions with
less contractible outputs will thus be more inclined to
attribute the formal use of the performance evaluation sys-
tem to their supervisor’s inherent and costly inclination to
provide accurate performance feedback.
In sum, we expect that the relationship between the
formal use of the performance evaluation system and trust
is different for different levels of contractibility. We expect
that contractibility affects the extent to which feedback
quality and justice are considered signs of trustworthiness.
In particular, our expectation can be summarized as
follows:
H5: The positive effect of the formal use of the perfor-
mance evaluation system on perceived performance
feedback quality and justice, and consequently on trust,
will be stronger for managers in functions with less
contractible outputs.
The proposed hypotheses are summarized in the model
in Fig. 1.
Method
To test these hypotheses, we conducted a survey study
among 160 departmental managers in eleven Slovenian
commercial banks. The choice of a survey study is based
on two considerations. First, there are no publicly available
archival data on the constructs that matter for our study.
Second, studies into justice and trust are usually perceived
to concern private information, which requires that the
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 727
data collection be anonymous. This is easily achieved by
the survey method. The choice for the banking sector
was based on the homogeneity of the sector, the compara-
bility of ?rms and, most importantly, the variation in con-
tractibility of outputs between managers.
To select survey respondents, we approached HR-man-
agers of all 18 Slovenian banks at the board level through
university contacts. Boards of directors of 11 banks agreed
to participate in the research. With the help of the human
resource board members, respondents in each bank were
selected based on the following three criteria. Managers
had to have at least one subordinate. They had to be in
their current position for at least one year to guarantee
their familiarity with the performance evaluation system.
Their compensation had to be contingent on performance,
and they had to experience at least one evaluation cycle, to
make sure that the evaluation system was a salient feature
of their working environment. Following these criteria a
total of 260 managers were selected for participation by
the banks’ HR-departments. There were a total of 160
usable responses, which represents a 61.5% response rate.
Table 3 reports on the number of managers for each of
the ?rms.
The survey instrument was web-based. We obtained e-
mail addresses of all respondents, and sent an individual-
ized e-mail message to each respondent with the request
to participate, a reference to the endorsement from the
bank, a unique and anonymous code, and a link to a web-
page of one of the sponsoring universities. Respondents
could login to the web-based questionnaire using the code.
In the design of the survey instrument, available guidelines
were followed (Dillman, 2000). Several measures were ta-
ken to increase the overall quality of the instrument, as
well as the response rate. First, the website was designed
by an experienced designer, and was made attractive by
using a limited number of colors, and an overall attractive
layout. Second, when possible, we used items to measure
our constructs from the extant literature. Third, the full
questionnaire was drafted in English, after which it was
translated into Slovenian using a translate–retranslate pro-
cedure. Four, the Slovenian questionnaire was subse-
quently tested with ?ve bank managers from three
different banks to assess the length, the understandability,
the attractiveness, and to search for any further dif?culties
respondents might have. In this phase small adaptations
were made to the wording of some questionnaire items
and the structure of the web instrument. This version of
the questionnaire was also intensively discussed with the
HR-managers. Finally, after sending the e-mails, a follow-
up procedure was used in which respondents were re-
minded two times by their employing organization based
on the codes provided by the researchers. Tables 1–3 con-
tain descriptive information on the sample of respondents.
The number of observations per organization ranged
from 4 to 35, re?ecting the organization’s size and respon-
sibility center structure (see Table 3).
Variable measurement
We measure formality of the performance evaluation
system (FORM) as a latent construct, for which we used
three manifest variables re?ecting each of the three steps
in the performance evaluation cycle (i.e., target setting,
performance measurement and rewarding) separately (cf.
Ittner & Larcker, 2001; Otley, 1999). We developed the
measurement instrument for formality of different subsys-
tems as follows.
Formality of target setting (FormTS) is measured using
four potential dimensions of performance in accordance
with the balanced scorecard logic used in Ittner, Larcker,
and Meyer (2003, see also the Appendix). We assessed
the relevance of these four-dimensions for our sample of
bank managers, as well as the examples of the perfor-
mance metrics in each category during the interviews in
the pre-test phase. We accounted for potential variation
of formality across dimensions and for the relative impor-
tance of each dimension by adopting the following mea-
surement procedure. First, we measured on ?ve-point
Likert scales (i) the extent to which performance targets
are explicitly documented in a written form and (ii) the ex-
Table 1
Sample descriptives.
Mean
(median)
Standard
deviation
Range Valid n
Age 43.45 (44.00) 8.18 29–58 153
Years in bank 9.82 (8.58) 8.58 1–35 157
Tenure in function 4.76 (3.25) 4.73 1–30 158
Table 2
Respondents’ areas of responsibility.
Job Subgroup Frequency Percent Cumulative
Trading Front-of?ce 59 36.6 38.1
Back-of?ce Back-of?ce 43 26.7 65.8
Investment banking
and asset
management
Front-of?ce 8 5.0 71.0
Treasury Front-of?ce 4 2.5 73.5
Administration Back-of?ce 35 21.7 96.1
International operations Front-of?ce 4 2.5 98.7
Other Back-of?ce 2 1.2 100.0
Missing n.a. 5 3.7
Total 160
Table 3
Participating banks and frequency of respondents.
Bank Frequency Percent Cumulative
Bank 1 13 8.1
Bank 2 21 13.1 8.1
Bank 3 35 21.9 21.3
Bank 4 15 9.4 43.1
Bank 5 11 6.9 52.5
Bank 6 13 8.1 59.4
Bank 7 9 5.6 67.5
Bank 8 14 8.8 73.1
Bank 9 20 12.5 81.9
Bank 10 5 3.1 94.4
Bank 11 4 2.5 97.5
Total 160 100.0 100.0
728 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
tent to which targets are quanti?ed, for each dimension
separately. We also asked respondents to allocate 100
points to the four dimensions to re?ect the importance of
a certain performance dimension for the respondent’s area
of responsibility. These weights were used to obtain a ?nal
score for formality of target setting.
Formality of performance measurement (FormPM) is
measured in accordance with the procedures outlined for
formality of target setting. We measure on ?ve-point Likert
scales (i) the extent to which the superior relies on objec-
tive information from the information system and (ii) the
extent to which respondent’s evaluation is expressed in
quantitative terms. Scores are obtained for the four perfor-
mance dimensions, which are weighted as described
before.
Formality of rewarding (FormREW) is measured with
items that address the objectivity of reward determination,
using ?ve-point Likert scales. The questions address
whether rewards are based on the superior’s judgment,
as opposed to objective information from the information
system, and whether rewards are based on a quantitative
as opposed to qualitative link with performance. These
questions were answered for both the ?xed and the vari-
able part of managerial rewards.
The two mediating variables are measured as follows.
Feedback quality of the supervisor (FEED) measurement
is based on the list of indicators developed by Steelman
et al. (2004), for which we used ?ve-point Likert scales. A
sample item is ‘My supervisor gives me useful feedback
about my job performance’. Procedural justice (PROCJ) is
measured with four items that address the extent to which
respondents believe that the subsystems of target setting,
performance measurement and rewarding, as well as the
system in total, leads to a fair determination of pay. A sam-
ple item is ‘I have full con?dence in the system’s fairness in
evaluating performance’. This particular focus of the
questions re?ects that our study investigates subjective
perceptions of procedural justice, in contrast with the vast
majority of studies that address either the antecedents or
effects of procedural justice (cf. Cropenzano, Byrne,
Bobocel, & Rupp, 2001). Thus, our questions are not based
on any normative model of justice (cf. Leventhal et al.,
1980), to avoid testing a tautology.
6
Trust in supervisor (TRUST) is measured by the instru-
ment developed by Read (1962). It addresses trust by ask-
ing subordinates about the extent to which they perceive
that their superior is inclined to take actions that are in
the interest of the subordinate. These questions were se-
lected as they are neutral vis-à-vis speci?c causes of trust,
and directly assess elements of trustworthiness. A sample
item is ‘My superior will always act in my favor if he has
the chance’.
Contractibility of managerial outputs was proxied by
the function of the manager in the bank. We distinguished
between two groups of managers according to their job
characteristics. The ?rst group consists of managers with
jobs that involve transactions that have direct ?nancial
consequences for the bank, which are measurable and ver-
i?able. These managers are in charge of front-of?ce depart-
ments that grant loans, hire deposits and other forms of
loans, manage assets, do transactions on foreign markets
and similar. The second group of managers provide sup-
port to the ?rst group in terms of expertise, analysis, con-
trol and administration. These activities include the
analyses of projects, ful?lling reporting requirements,
internal audit, legal counselling and internal administra-
tion. These back-of?ce functions, while crucial for bank
performance, have less measurable and veri?able outputs,
and their effect on bank performance is less traceable.
Thus, for performance evaluation, this distinction seems
to capture the essential difference between managerial
performance of the ?rst group that is immediately and di-
rectly re?ected on the bank’s market performance, and the
performance effects of the second group’s activities which
are less obvious, and only indirectly and imperfectly linked
to market transactions.
We included two control variables which are tenure
(TENURE) and task uncertainty (UNCERT). The former is
measured with number of years the respondent has been
on the present managerial position. It controls for the trust
effects of tenure (cf. Gibbs et al., 2004). The latter is mea-
sured with the instrument developed by Withey, Daft,
and William (1983). This variable controls for uncertainty’s
potential effect on the use of performance measures.
Results
We tested our hypotheses by using a partial least
squares (PLS) method for structural equation model esti-
mation SEM (Lohmöller, 1989; Wold, 1982; Wold,
1985). Contrary to the variance–covariance-based method
of SEM estimation that attempts to minimize the differ-
ence between sample covariances and those predicted
by a theoretical model on the underlying assumptions
of normal multivariate distribution (Chin, 1998), PLS esti-
mates the structural model using an iterative OLS regres-
sion-like procedure, which aims to explain variance of the
dependent variables by minimizing the residual variance
of all dependent variables (both, latent and observed). Be-
cause of its iterative estimation algorithm that proceeds
block-by-block, PLS requires less stringent assumptions
about the distributional characteristics of the raw data
and sample size. We used SmartPLS software version
2.00.
Before the assessment of the structural model, we as-
sessed the quality of the measurement model, thus
addressing individual item reliability, construct reliability,
and convergent and discriminant validity for our re?ective
constructs (cf. Bagozzi, 1994). Individual item reliability is
considered adequate when an item has a factor loading
that is greater than 0.71 on its respective construct which
implies that more than 50% of the variance in the observed
variable is shared with the construct (cf. Chin, 1998). The
measure of internal consistency or construct reliability is
6
Several studies in the procedural justice literature apply a normative
model of justice that prescribes what process characteristics make a
process ‘fair’ (e.g., Taylor et al., 1998). Measuring the perceived presence of
these characteristics would be tautological as a procedural element would
then be an independent and a dependent variable at the same time.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 729
Dillon-Goldstein q (Tenenhaus, Vinzi, Chatelin, & Lauro,
2005). This measure is preferable to the Cronbach’s alpha,
as the latter assumes that all loadings are equal to one. A
threshold of 0.70 is suggested as a modest level of reliabil-
ity (Vandenbosch, 1996). Convergent and discriminant
validity of the constructs is assessed by the average vari-
ance extracted (AVE), which represents the average vari-
ance shared between a construct and its indicators
(Fornell & Larcker, 1981). For convergent validity, AVE
should be greater than 0.50. This ensures that measure-
ment error does not dominate the variance captured by
the construct (Vandenbosch, 1996). To assess discriminant
validity taking into account model parameters and mea-
surement errors, AVE should be greater than the variance
shared between any two constructs and the cross-loadings
of indicators on other constructs should be minimal.
The second step in the PLS analysis is the estimation of
the speci?ed structural equations. The path coef?cients
indicate the strength and direction of the relationships
among the latent variables. We assessed statistical signi?-
cance of parameter estimates using a bootstrap procedure
with 1000 replacements. In addition, the predictive valid-
ity of the parameter estimates can be assessed via a
cross-validated redundancy index or a so called Stone–
Geisser Q
2
-test (Geisser, 1974; Stone, 1974). A result indi-
cating that Q
2
for all latent variables is greater than 0 sug-
gests that the model has predictive relevance. As PLS
models lack an index that can provide the goodness of ?t
statistics like in variance–covariance-based SEM-ML mod-
els, Tenenhaus et al. (2005) and Vandenbosch (1996) argue
that next to the reliability and validity of constructs, the
signi?cance of variance explained and positive Q
2
s for all
constructs provide suf?cient evidence of model ?t.
The results of the measurement model for the full sam-
ple are summarized in Tables 4–6. As Table 4 reports, we
established that all factor loadings were greater than 0.71
with the exception of one indicator of TRUST. This provides
evidence of a satisfactory level of individual item reliabil-
ity. The indicator for TRUST, for which the loading of 0.69
was obtained, was maintained in the analysis given the
satisfactory construct reliability measured by q
c
, which ex-
ceeded 0.83 in all cases. Convergent validity appeared
acceptable for all re?ective constructs, as AVE exceeded
0.5 in all cases. Discriminant validity of constructs is pre-
sented in Table 5. It shows that diagonal AVE values exceed
all other scores, suggesting suf?cient discriminant
validity.
7
We performed our analysis of the hypotheses in a step-
wise fashion (Baron & Kenny, 1986; Luft & Shields, 2003),
such that we ?rst tested whether formality affected trust
directly (hypothesis 1, see Table 7, direct relationship mod-
el), and how this relationship altered by the subsequent
introduction of our two mediators according to hypothesis
2 and 3 (see Table 7, partially mediated relationship and
fully mediated relationship model). The results of the anal-
ysis for the full sample are presented in Table 7 panel A and
signi?cant (positive) paths are depicted as solid lines in
Fig. 2. The path coef?cient from FORM to TRUST shows
support for hypothesis 1, as FORM has a positive and statis-
tically signi?cant direct impact on TRUST (0.247, p < 0.001,
see direct relationship model). This effect is reduced when
PROCJ is introduced into the model, although the results do
not fully support hypothesis 2 as only the relationship be-
tween PROCJ and TRUST is signi?cant (0.20, p < 0.001),
whereas the direct path between FORM and PROCJ is insig-
ni?cant. When FEED is introduced into the model (hypoth-
esis 3) the direct link between FORM and TRUST is further
reduced and becomes insigni?cant, displaying a full medi-
ation effect (cf. Baron & Kenny, 1986). Both variables now
absorb the explanatory power of the initial direct effect of
FORM on TRUST. For FEED, the path coef?cient to TRUST is
0.412 (p < 0.001). The path coef?cient from FORM to FEED
is positive and signi?cant too (0.254, p < 0.01), which sup-
ports hypothesis 3. In support of hypothesis 4, FEED also
mediates the relationship between FORM and PROCJ as
the path coef?cient of FEED to PROCJ is 0.413 (p < 0.001).
Note that these results are obtained controlling for tenure
and task uncertainty.
Hypothesis 5 expressed our prediction that the relation-
ships between FORM and the two mediating variables
would be different for managers with more and less con-
tractible outputs. To test this hypothesis, we re-estimate
both the measurement model and the structural model
for the two subgroups of managers in front-of?ce and
back-of?ce functions, that we used as a proxy for high
and low contractibility of managerial outputs, respectively
(cf. Grossman & Hart, 1986). Table 2 shows how managers
in different areas of responsibility were allocated to the
subgroup analysis. Measurement model and reliability/
validity analyses for these subgroups are not presented
here for brevity purposes, but show no material differences
with the presented results for the full sample. The path
models for the front-of?ce subgroup and the back-of?ce
subgroup are presented in Table 7, panels B and C, respec-
tively. Figs. 3 and 4 depict the signi?cant positive paths for
the two subgroups.
The results suggest that for the high-contractibility sub-
group, formality of performance evaluation has no effect
on trust in the direct relationship model, nor is it linked
to procedural justice and feedback quality in the mediated
relationship model. For the low-contractibility subgroup,
the results are largely in line with the results obtained
for the full sample. Formality appears to affect trust (direct
relationship model), but the relationship appears to be
mediated by both procedural justice and feedback quality
(fully mediated relationship model). In the indirect model,
feedback quality also appears to mediate the relationship
between formality and procedural justice, as was expected.
In line with the results for the full sample, this suggests
that for the back-of?ce subgroup, feedback quality is the
crucial variable in the formality–trust relationship. As ex-
7
When more items per latent construct were available, we used
exploratory factor analysis using SPSS with Varimax rotation to identify
the best indicators for use in the measurement model. After eliminating
nine items on the basis of their low loadings on the respective factors (Hair,
Anderson, Tatham, & Black, 1995), we retained 14 indicators for the four
re?ective constructs (see the Appendix for an overview of the latent
constructs and their indicators). We performed a test of response bias by
comparing average scores on all variables for early and late respondents,
but found no evidence of response bias above chance (Blair & Zinkhan,
2006). The questionnaire is available from the corresponding author.
730 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
pected the same relationship is insigni?cant in the high-
contractibility subgroup. Such, too, is the relationship be-
tween procedural justice and trust in this subgroup. This
suggests that for the front-of?ce subgroup, formality of
performance evaluation does not invoke subordinate’s
trust of the superior, nor that higher formality would pos-
itively impact the perception of procedural justice or qual-
ity of feedback. As shown in Table 8, the differences in
strength of formality on feedback quality and perceptions
of justice between the two subgroups are signi?cant.
8
Discussion and conclusions
The purpose of this paper is to explore the effects of
superiors’ use of the performance evaluation system on
Table 4
Estimation of the measurement model parameters (full sample, n = 160).
Loading original sample Loading sample mean Standard deviation t-Statistic (bootstrap) Composite reliability
(q
c
)
Average variance
extracted (AVE)
FORM 0.882 0.715
FormTS 0.837 0.831 0.060 13.832
FormPM 0.793 0.785 0.074 10.730
FormRew 0.904 0.899 0.057 15.797
FEED 0.903 0.700
Feed
1
0.805 0.806 0.030 26.469
Feed
2
0.858 0.853 0.035 24.774
Feed
3
0.888 0.884 0.027 33.487
Feed
4
0.793 0.791 0.045 17.542
PROCJ 0.910 0.718
Procj
1
0.784 0.784 0.044 17.838
Procj
2
0.838 0.835 0.041 20.201
Procj
3
0.899 0.898 0.020 44.558
Procj
4
0.864 0.864 0.027 31.944
TRUST 0.845 0.647
Trust
1
0.847 0.846 0.028 30.219
Trust
2
0.871 0.869 0.025 34.501
Trust
3
0.681 0.676 0.080 8.507
UNCERT 0.902 0.754
Uncert
1
0.897 0.873 0.146 6.143
Uncert
2
0.882 0.856 0.149 5.903
Uncert
3
0.823 0.795 0.137 6.014
TENURE n.a. n.a.
Table 5
Discriminant validity coef?cients (full sample, n = 160).
FORM TENURE UNCERT PROCJ FEED TRUST
FORM 0.845
TENURE À0.140 n.a.
UNCERT 0.179 À0.008 0.868
PROCJ 0.172 À0.034 0.280 0.883
FEED 0.254 À0.160 0.163 0.430 0.836
TRUST 0.274 0.014 0.265 0.408 0.523 0.804
Diagonal elements (bold) are the square root of the variance shared
between the constructs and their indicators (AVE). Off-diagonal elements
are the correlations among constructs. For discriminant validity, diagonal
elements should be larger than off-diagonal elements. n.a. is non-
applicable.
Table 6
Cross-loadings (full sample, n = 160).
FORM FEED TRUST PROCJ UNCERT TENURE
FORM
FormPM 0.837 0.168 0.107 0.060 0.085 À0.117
FormTS 0.793 0.148 0.201 0.070 0.133 À0.034
FormRew 0.904 0.279 0.309 0.228 0.195 À0.170
FEED
Feed
1
0.226 0.805 0.480 0.448 0.152 À0.150
Feed
2
0.178 0.858 0.413 0.321 0.130 À0.113
Feed
3
0.207 0.888 0.443 0.365 0.127 À0.077
Feed
4
0.237 0.793 0.400 0.277 0.132 À0.200
TRUST
Trust
1
0.209 0.475 0.847 0.317 0.225 0.069
Trust
2
0.237 0.499 0.871 0.409 0.249 À0.024
Trust
3
0.232 0.218 0.681 0.229 0.147 À0.021
PROCJ
Procj
1
0.120 0.257 0.314 0.784 0.171 À0.005
Procj
2
0.157 0.464 0.342 0.838 0.250 À0.055
Procj
3
0.165 0.372 0.374 0.899 0.308 0.009
Procj
4
0.133 0.333 0.351 0.864 0.205 À0.059
UNCERT
Uncert
1
0.178 0.120 0.240 0.228 0.897 0.004
Uncert
2
0.167 0.196 0.283 0.315 0.882 À0.051
Uncert
3
0.106 0.097 0.141 0.166 0.823 0.041
TENURE
Tenure
1
À0.140 À0.160 0.014 À0.034 À0.009 1.000
8
We also analyzed the signi?cance of the difference in path-coef?cients
for the direct model between the two subgroups. These coef?cients were
0.026 and 0.370 for the front-of?ce and back-of?ce subgroups, respectively,
with the difference signi?cant at p < 0.05. This con?rms the outcomes of
the subgroup analysis for the mediated relationship model.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 731
subordinates’ trust in their superiors. The extant litera-
ture shows con?icting results when investigating the
relationship between formal control systems and trust
for collaborative team arrangements in general, and for
the relationships between superior and subordinate in
particular. We focused on superiors’ formal use of the
performance evaluation system, which we argued to be
an important underlying, generic aspect of the use of
such systems, and set out to explore the relationship be-
formality trust
feedback
quality
procedural
justice
Fig. 3. Results of the mediation model (front-of?ce subgroup, n = 75,
signi?cant paths displayed).
Table 7
Direct and indirect effects.
Path Direct relationship Partially mediated relationship Fully mediated relationship Q
2
R
2
Panel A: full sample, n = 160
a
Effect on TRUST 0.160 0.358
FORM > TRUST 0.247
***
0.208
**
0.148
PROCJ > TRUST 0.333
***
0.209
**
FEED > TRUST 0.412
***
Effect on FEED 0.044 0.080
FORM > FEED 0.254
**
Effect on PROCJ 0.127 0.233
FORM > PROCJ 0.130 0.066
FEED > PROCJ 0.413
***
Panel B: front-of?ce subsample (n = 75)
Effect on TRUST 0.162 0.325
FORM > TRUST 0.026 0.026 À0.006
PROCJ > TRUST 0.126 À0.035
FEED > TRUST 0.490
***
Effect on FEED 0.013 0.018
FORM > FEED 0.052
Effect on PROCJ 0.103 0.164
FORM > PROCJ 0.036 0.015
FEED > PROCJ 0.272
**
Panel C: back-of?ce subsample (n = 80)
Effect on TRUST 0.159 0.432
FORM > TRUST 0.370
***
0.284
**
0.208
PROCJ > TRUST 0.446
***
0.314
**
FEED > TRUST 0.343
***
Effect on FEED 0.071 0.105
FORM > FEED 0.314
**
Effect on PROCJ 0.141 0.213
FORM > PROCJ 0.222 0.103
FEED > PROCJ 0.397
***
**
p < 0.01 (based on t
(1000)
, two-tailed test).
***
p < 0.001 (based on t
(1000)
, two-tailed test).
a
The number of observations in the full sample (160) exceeds the sum of two subsamples (155) as ?ve managers have not reported their function and are
excluded from the subsample analysis.
formality
trust
feedback
quality
procedural
justice
Fig. 2. Results of the mediation model (full sample, n = 160, signi?cant
paths displayed).
formality trust
feedback
quality
procedural
justice
Fig. 4. Results of the mediation model (back-of?ce subgroup, n = 80,
signi?cant paths displayed).
732 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
tween formality of performance evaluation and interper-
sonal trust. Our ?ndings overall suggest that formality of
performance evaluation system use is indeed a relevant
characteristic in evoking subordinate trust. This ?nding
itself may explain why previous studies have found trust
to be enhanced by both ?nancial and non-?nancial per-
formance criteria that are both re?ections of formality
(Lau & Buckland, 2001; Lau & Shohilin, 2005), as both
suggest a formal use of metrics, rather than informal
evaluations (cf. Moers, 2005). For the overall model the
results con?rm the hypotheses that the link between for-
mality of the performance management cycle and trust is
mediated through the quality of feedback and procedural
justice (see Fig. 2). In particular, we ?nd that formality
matters for trust because it enhances the perceived
quality of feedback for subordinates. We found that high-
er perceived quality of feedback results in higher trust,
both directly, and via higher perceptions of procedural
justice.
In addition, we ?nd that the relationship between for-
mality of performance evaluation and trust is different
for managers in different situations, as attribution theory
predicts. We use the term contractibility to denote
differences in functions concerning easiness of de?ning
measurable outputs, and proxied for this variable by dis-
tinguishing front-of?ce managers, whose main responsi-
bilities regard market transactions, from back-of?ce
managers, whose main responsibilities regard internal
services. Managers in functions with highly contractible
outputs seem to be less affected by the level of formality
that their superiors apply in the performance evaluation
process than managers in functions with less measurable
outputs. These managers will see fair and transparent
performance feedback less as an outcome of their supe-
riors’ formal use of the system, but rather as a conse-
quence of the nature of their function and the
performance evaluation system (cf. Herold & Parsons,
1985). As their functional output itself provides formal
feedback immediately, little is attributable to the supe-
rior’s use of the PE system since superiors need little ef-
fort to formally use the available performance
information in their evaluation process (Sitkin & George,
2005). Managers in back-of?ce function, however, seem
affected by the formality that superiors apply in the per-
formance evaluation process. This means that the effort
superiors put in operationalizing output which by itself
is not easily measurable and traceable, results in trust
attributions.
For back-of?ce managers instead, formality has a po-
sitive effect on trust through both its effect on feedback
and justice. Feedback quality, however, appears to be
the most important factor in terms of the mediation ef-
fects explored. This implies that the link between perfor-
mance evaluation and trust is more complex than is
currently recognized in the procedural justice literature
that claims a central role for justice (e.g., Cohen-Charash
& Spector, 2003, p. 283). Indeed, when outputs are less
measurable, managers seem to perceive formal evalua-
tions, in which a superior puts less emphasis on his or
her subjective judgments, as more just. Our analysis
shows, however, that this is an indirect effect, which is
caused by the impact that improved performance
feedback has on justice perceptions. This is an important
?nding that suggest that studies on the relationship be-
tween performance evaluation systems and trust should
refocus on the feedback role of these systems (e.g., Lau
& Buckland, 2001; Lau & Shohilin, 2005). A ?nal differ-
ence between the two subgroups relates to the rele-
vance of justice for trust. For the front-of?ce managers,
justice and trust were unrelated, suggesting that other
superior’s behaviors, such as providing high-quality
feedback outside the performance evaluation cycle,
drives trust. We did not ?nd a signi?cant negative effect
of performance evaluation system use on trust as some
earlier studies reported. In these studies these negative
effects of controls on trust were found when controls
were newly installed (e.g., Tenbrunsel & Messick,
1999), and when existing formal and informal controls
were abolished (Malhotra & Munighan, 2002). Our study
instead focused on the continuous use of an existing set
of controls (i.e., the performance evaluation system),
applying a cross-sectional rather than longitudinal
analysis.
Finding differences between managerial groups is
important, since it not only suggests that the antecedents
of trust may be different across organizational functions,
but also explains previous contradictory results. First,
our ?nding of differences in trust attributions between
our two subgroups provides some reconciliation to the
opposite ?ndings of Coletti et al. (2005) and Malhotra
and Munighan (2002). Our study shows how context
may affect these attributions, which means that the nat-
ure or tasks may explain differential attributions in their
experimental settings. Note that both, in the earlier ?eld
studies (e.g., Lau & Buckland, 2001; Lau & Shohilin,
2005) and in the experimental studies (Coletti et al.,
2005; Tenbrunsel & Messick, 1999) research designs have
Table 8
Signi?cance of path differences.
a
Mediation model Front-of?ce Back-of?ce Difference T-test
Effect on TRUST
FORM > TRUST À0.006 0.208 À0.213 n.s.
PROCJ > TRUST À0.035 0.314
**
À0.349 À2.133
*
FEED > TRUST 0.490
***
0.343
***
0.147 n.s.
Effect on FEED
FORM > FEED 0.052 0.314
**
À0.262 À1.660
*
Effect on PROCJ
FORM > PROCJ 0.015 0.103 À0.088 n.s.
FEED > PROCJ 0.272
**
0.397
***
À0.125 n.s.
n.s., not signi?cant, df = 153.
*
p < 0.05 (two-tailed test).
**
p < 0.01 (two-tailed test).
***
p < 0.001 (two-tailed test).
a
Corresponding path coef?cients have been compared across back-
and front-of?ce using the following test t ¼
Path
fo
ÀPath
bo
???????????????????????????????????????????
ðmÀ1Þ
2
ðmþnÀ2Þ
ÂS:E
2
fo
þ
ðnÀ1Þ
2
ðmþnÀ2Þ
ÂS:E
2
bo
q h i
Â
???????
1
m
þ
1
n
p  Ã
where m is the sample size of front-of?ce,
n is the sample size of back-of?ce, fo is the front-of?ce, bo is the back-
of?ce, with m + n À 2 degrees of freedom.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 733
not explicitly taken into account the contractibility of
managerial outputs that we addressed through our sub-
group analyses.
9
This study is one of the few studies that investigate
the role of formal control systems in creating trust,
and one of the ?rst studies looking at hierarchical col-
laborative setting of the superior and the subordinate.
Obviously, the study is vulnerable to the typical weak-
nesses of survey research, regarding validity and reliabil-
ity of items and tests, although in the design and
execution we have avoided many of the potential pit-
falls. Both in the development and pre-test of the instru-
ment, and in the approach of respondents we followed
best practice. For some of the variables that we devel-
oped for this study, future studies may establish their
wider usefulness and validity. One ?nal potential limita-
tion of this study lies in our sample, in which we cannot
exclude the possibility that several subordinates have re-
ferred to the same superior in their answers. Although
this may have reduced the variance in scores, and there-
fore may have caused some bias, we believe reduced
variance in fact has decreased rather than increased
our chances of ?nding non-existent effects. Therefore,
we believe that our results cannot be explained by this
effect.
Future studies may look into the relationships between
performance evaluation systems and trust in further de-
tail. Clearly, it would be fruitful to replicate our survey
study using an experimental design, which is the method
used in much of the prior research. Such a design would
allow manipulating the level of contractibility across
experimental conditions, and con?rm the validity of our
proxy. It would also add a formal test of causality. Repli-
cation studies in other industries and functions would al-
low the asses the generalizability of our results. The study
provides numerous directions for future research. Future
studies could extend our study by analyzing effects of
other combinations of performance measurement
arrangements, and external conditions on trust. Overall,
however, we feel that the apparent complexity and rele-
vance of trust in the superior–subordinate setting pro-
vides a fruitful avenue for studies in performance
evaluation and control.
Acknowledgements
We thank participants at the annual EAA-conference,
the semi-annual AAA management accounting conference,
Kristy Towry, Ken Merchant, Henri Dekker, David Emsley,
Mike Shields, Simon C
?
adez? and seminar participants at
RSM Erasmus University, the University of
Ljubljana, and the University of Örebro for their comments
and suggestions towards earlier versions of this paper.
Appendix
Measurement instrument
Formality of performance evaluation system (FORM)
Question 1: Weighting of performance dimensions
This question is about the responsibilities of your job as
manager of a unit (e.g., a branch, or a department).
Below you ?nd four ways in which your supervisor may
assess your performance as unit manager. We are
interested in the relative importance of each of these
four performance areas. Therefore we ask you to allocate
a total number of 100 points to one or more of these
areas, re?ecting their relative importance. For example,
if you allocate the full 100 points to one area, you
express that only that area is important, and not the
other 3. If you allocate 25 points to each area, you
indicate that all areas are equally important for your
work, and that your superior does not have any priorities
or emphases. Please be as speci?c as possible in
allocating the 100 points, and answer these questions for
your function as unit manager, rather than for the bank
in general. There are no wrong or right answers. Note
that the areas are described broadly
Financial performance of your unit (e.g., sales,
pro?tability, interest margins, costs, value of
collected deposits, value of granted loans, value
of payment commissions, etc.)
Operational performance of your unit (e.g.,
productivity, completion of projects and phases
of projects, number of activities such as issued
credit cards, number of stipulated
contracts, etc.)
Employee relations within your unit (e.g., employee
satisfaction, turnover, achieved workforce
capabilities, number of trainings, etc.)
Quality and innovation of your unit (e.g., quality
awards, improvement of service quality,
customer satisfaction, introduction
of new services, ef?ciency
improvements, etc.)
Total (points or %) 100%
Question 2: Formality of target setting (FormTS)
Level of agreement with the following statements:
Less formal Scale More formal
When establishing
work objectives, my
superior does not
explicitly document
these objectives in
writing
1 2 3 4 5 When establishing
work objectives, my
superior explicitly
documents these
objectives in writing
When establishing
work objectives, my
superior casts these
objectives in
qualitative terms
1 2 3 4 5 When establishing
work objectives, my
superior casts these
objectives in
quantitative targets
9
The tasks that experimental subjects performed in the experimental
studies mentioned (Coletti et al., 2005; Malhotra & Munighan, 2002;
Tenbrunsel & Messick, 1999) differed widely in terms of complexity and
realism. As task design was not a treatment in any of these experimental
studies, explaining differences in trust attributions through task design
would be mere speculation.
734 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
This question was asked for all areas of performance, and
scores were averaged using the weights from question 1.
Question 3: Formality of performance measurement
(FormPM)
Level of agreement with the following statements:
Less formal Scale More formal
When judging my
performance, my
superior uses his
(her) personal
judgment of my
performance
1 2 3 4 5 When judging my
performance, my
superior relies on
objective information
from the information
system
When judging my
performance, my
superior discusses
my performance in
qualitative terms
1 2 3 4 5 When judging my
performance, my
superior expresses my
performance in
quantitative terms
(rates my performance
on quantitative rating)
This question was asked for all areas of performance, and
scores were averaged using the weights from question 1.
Question 4: Formality of rewarding (FormRew)
Level of agreement with the following statements:
Less formal Scale More formal
My ?xed pay is based on
my superior’s
personal judgment of
my performance
1 2 3 4 5 Objective
information from the
information system
My ?xed pay is based on
my performance in
qualitative terms
1 2 3 4 5 My performance in
quantitative terms
My bonus is based on
my superior’s
personal judgment of
my performance
1 2 3 4 5 Objective
information from the
information system
My bonus is based on
my performance in
qualitative terms
1 2 3 4 5 My performance in
quantitative terms
Question 5: Feedback quality (FEED)
Level of agreement with the statements (1 = I com-
pletely disagree, 2 = I disagree, 3 = neutral, 4 = I agree and
5 = I completely agree):
Feed
1
My superior gives me useful feedback about my job
performance
Feed
2
The performance information I get from my
superior is generally not very meaningful (reverse)
Feed
3
The feedback I receive from my superior helps me
do my job
Feed
4
I value the feedback I receive from my superior
Question 6: Procedural justice (PROCJ)
Level of agreement with the statements (1 = I com-
pletely disagree, 2 = I disagree, 3 = neutral, 4 = I agree and
5 = I completely agree):
Procj
1
I have full con?dence in the system’s fairness in
determining pay
Procj
2
I have full con?dence in the system’s fairness in
determining targets
Procj
3
I think that the way in which my performance is
measured is fair
Procj
4
I have full con?dence in the system’s fairness in
evaluating performance
Question 7: Trust in supervisor (TRUST)
Level of agreement with the statements (1 = I com-
pletely disagree, 2 = I disagree, 3 = neutral, 4 = I agree and
5 = I completely agree):
Trust
1
My superior will always act in my favor if he has
the chance
Trust
2
I am convinced that my superior will always fully
and honestly keep me up to date of everything
that is important to me
Trust
3
If my superior takes a decision that is against my
interest, I am convinced that this decision is
justi?ed for other reasons
Question 8: Task Uncertainty (UNCERT)
Level of agreement with the statements (1 = I com-
pletely disagree, 2 = I disagree, 3 = neutral, 4 = I agree and
5 = I completely agree):
Uncert
1
There is a clearly known way to do the major
types of work I normally encounter (reverse)
Uncert
2
There is a clearly de?ned body of knowledge of
subject matter which can guide me in doing my
work (reverse)
Uncert
3
There is an understandable sequence of steps
that can be followed in doing my work (reverse)
References
Bagozzi, R. P. (1994). Structural equation models in marketing research:
Basic principles. In Bagozzi (Ed.), Principles of marketing research.
Blackwell Business.
Birnberg, J., Frieze, I., & Shields, M. D. (1977). The role of attribution theory
in control systems. Accounting, Organizations and Society, 3, 189–200.
Blair, E., & Zinkhan, G. (2006). Nonresponse and generalizability in
academic research. Academy of Marketing Science, 34(1), 4–7.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable
distinction in social psychological research: Conceptual, strategic and
statistical considerations. Journal of Personality and Social Psychology,
51, 1173–1182.
Cahuc, P., & Zylberberg, A. (2004). Labour economics. Cambridge, MA: The
MIT Press.
Chenhall, R. H., & Lang?eld-Smith, K. (2003). Performance measurement
and reward systems, trust, and strategic change. Journal of
Management Accounting Research, 15, 117–143.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 735
Chin, W. W. (1998). The partial least squares approach to structural
equation modeling. In G. A. Marcoulides (Ed.), Modern methods for
business research (pp. 295–336). Mahwah, NJ: Lawrence Erlbaum
Associates, Publisher.
Clark, M. C., & Payne, R. L. (1997). The nature and structure of workers’
trust in managers. Journal of Organizational Behavior, 18, 205–224.
Cohen-Charash, Y., & Spector, P. E. (2003). The role of justice in
organizations: A meta-analysis. Organizational Behavior and Human
Decision Processes, 86, 278–321.
Coletti, A. L., Sedatole, K. L., & Towry, K. L. (2005). The effect of control
systems on trust in collaborative environments. The Accounting
Review, 80, 477–500.
Colquitt, J. A., & Jackson, J. L. (2006). Justice in teams: The context
sensitivity of justice rules across individual and team contexts. Journal
of Applied Social Psychology, 36, 868–899.
Cropenzano, R., Byrne, Z. S., Bobocel, D. R., & Rupp, D. E. (2001). Moral
virtues, fairness heuristics, social entities and other denizens of
organizational justice. Journal of Vocational Behavior, 58, 164–209.
Das, T. K., & Teng, B. S. (1998). Between trust and control: Developing
con?dence in partner cooperation in alliances. Academy of
Management Review, 23, 491–512.
Dekker, H. (2004). Control of inter-organizational relationships: Evidence
on appropriation concerns and coordination requirements.
Accounting, Organizations, and Society, 29, 27–49.
Dillman, D. A. (2000). Mail and Internet surveys: The tailored design method
(second ed.). New York, NY: John Wiley & Sons.
Dirks, K. T., & Ferrin, D. L. (2002). Trust in leadership: Meta-analytic
?ndings and implications for research and practice. Journal of Applied
Psychology, 87, 611–628.
Erdogan, B. (2002). Antecedents and consequences of justice perceptions
in performance appraisals. Human Resource Management Review, 12,
555–578.
Fine, G., & Holy?eld, L. (1996). Secrecy, trust, and dangerous leisure:
Generating group cohesion in voluntary organizations. Social
Psychology Quarterly, 59, 22–38.
Fisher, J. G., Maines, L. A., Peffer, S. A., & Sprinkle, G. B. (2005). An
experimental investigation of employer discretion in employee
performance evaluation and compensation. The Accounting Review,
80, 563–583.
Fornell, C., & Larcker, D. (1981). Evaluating structural equation models
with unobservable variables and measurement error. Journal of
Marketing Research, 18, 39–50.
Free, C. (2008). Walking the talk? Supply chain accounting and trust
among UK supermarkets and suppliers. Accounting, Organizations and
Society, 33, 629–662.
Gambetta, D. (Ed.). (1988). Trust: Making and breaking cooperative
relations. Oxford: Basil Blackwell.
Geisser, S. (1974). A predictive approach to the random effect model.
Biometrika, 61, 101–107.
Gibbs, M., Merchant, K. A., Van der Stede, W. A., & Vargus, M. E. (2004).
Determinants and effects of subjectivity in incentives. The Accounting
Review, 79, 409–436.
Greenberg, J. (1986). Determinants of perceived fairness of performance
evaluations. Journal of Applied Psychology, 71, 340–342.
Grossman, S. J., & Hart, O. D. (1986). The costs and bene?ts of ownership:
A theory of vertical and lateral integration. Journal of Political
Economy, 94, 691–719.
Hair, J., Anderson, R. E., Tatham, R. L., & Black, W. C. (1995). Multivariate
data analysis (fourth ed.). Englewood Cliffs, NJ: Prentice-Hall.
Hardin, R. (1992). The street-level epistemology of trust. Analyse & Kritik,
Zeitschrift fuer Sozialtheorie, 14, 152–176.
Hartmann, F. G. H. (2000). The appropriateness of RAPM: Toward the
further development of theory. Accounting, Organizations and Society,
25, 451–482.
Hartmann, F. G. H. (2005). The effects of tolerance for ambiguity and
uncertainty on the appropriateness of accounting performance
measures. ABACUS, 41, 241–264.
Hartmann, F. G. H. (2007). Do accounting performance measures indeed
reduce managerial ambiguity under uncertainty? Advances in
Management Accounting Research, 16, 159–180.
Herold, D. M., & Parsons, C. K. (1985). Assessing the feedback environment
in work organizations: Development of the job feedback survey.
Journal of Applied Psychology, 70, 290–305.
Hopwood, A. G. (1972). An empirical study of the role of accounting data
in performance evaluation. Journal of Accounting Research, 10,
156–182.
Ittner, C. D., & Larcker, D. F. (2001). Assessing empirical research in
managerial accounting: A value-based management perspective.
Journal of Accounting and Economics, 32, 349–410.
Ittner, C. D., Larcker, D. F., & Meyer, M. (2003). Subjectivity and the
weighting of performance measures: Evidence from a balanced
scorecard. The Accounting Review, 78(3), 725–758.
Jones, T. M. (1995). Instrumental stakeholder theory: A synthesis of ethics
and economics. Academy of Management Review, 20, 404–437.
Kramer, R. M. (1999). Trust and distrust in organizations: Emerging
perspectives, enduring questions. Annual Review of Psychology, 50,
569–598.
Landy, F. J., Barnes, J. L., & Murphy, K. R. (1978). Correlates of perceived
fairness and accuracy of performance evaluation. Journal of Applied
Psychology, 63, 751–754.
Lau, C. M., & Buckland, C. (2001). Budgeting – Role of trust and
participation: A research note. ABACUS, 37, 369–388.
Lau, C. M., & Shohilin, M. (2005). Financial and non?nancial performance
measures: How do they affect job satisfaction. British Accounting
Review, 37, 389–413.
Lau, C. M., & Tan, S. L. C. (2006). The effects of procedural fairness and
interpersonal trust on job tension in budgeting. Management
Accounting Research, 17, 171–186.
Leventhal, G. S. (1976). Fairness in social relationships. In J. W. Thibaut, J.
T. Spence, & R. C. Carson (Eds.), Contemporary topics in social
psychology (pp. 211–239). Morristown, New York: General Learning
Press.
Leventhal, G. S., Karuza, J., & Fry, W. R. (1980). Beyond fairness: A theory
of allocation preferences. In G. Mikula (Ed.), Justice and social
interaction (pp. 167–218). New York: Springer-Verlag.
Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task
performance. Englewood-Cliffs: Prentice-Hall.
Lohmöller, J. B. (1989). Latent variable path modeling with partial least
squares. Heidelberg: Physica-Verlag.
Luft, J., & Shields, M. D. (2003). Mapping management accounting:
Graphics and guidelines for theory-consistent empirical research.
Accounting, Organizations and Society, 28, 169–249.
Malhotra, D., & Munighan, J. K. (2002). The effects of contracts on
interpersonal trust. Administrative Science Quarterly, 47, 534–559.
March, J. G., & Olsen, J. P. (1989). Rediscovering institutions. New York: Free
Press.
Marginson, D., & Ogden, S. (2005). Coping with ambiguity through the
budget: The positive effects of budgetary targets on managers’
budgeting behaviours. Accounting, Organizations and Society, 30,
435–456.
Moers, F. (2005). Discretion and bias in performance evaluation: The
impact of diversity and subjectivity. Accounting, Organizations and
Society, 30, 67–80.
Morand, D. D. (1995). The role of behavioral formality and informality in
the enactment of bureaucratic versus organic organizations. Academy
of Management Review, 20, 831–872.
Otley, D. T. (1999). Performance management: A framework for
management control systems research. Management Accounting
Research, 10, 363–382.
Read, W. H. (1962). Upward communication in industrial hierarchies.
Human Relations, 15, 3–15.
Ross, A. (1994). Trust as a moderator of the effect of performance
evaluation style on job-related tension: A research note. Accounting,
Organizations and Society, 19, 629–635.
Rousseau, D. M., Sitkin, S. B., Byrt, R. S., & Camerer, C. (1998). Not so
different after all: A cross-discipline view of trust. Academy of
Management Review, 23, 393–404.
Sitkin, S. B., & George, E. (2005). Managerial trust-building through the
use of legitimating formal and informal control systems. International
Sociology, 20, 307–338.
Steelman, L. A., Levy, P. E., & Snell, A. F. (2004). The feedback environment
scale: Construct de?nition, measurement and validation. Educational
and Psychological Measurement, 64, 165–184.
Stone, M. (1974). Cross validity choice and assessment of statistical
predictions. Journal of the Royal Statistical Society, Series B, 36,
111–133.
Taylor, M. S., Tracy, K. B., Renard, M. K., Harrison, J. K., & Carroll, S. J.
(1995). Due process in performance appraisal: A quasi-experiment in
procedural justice. Administrative Science Quarterly, 40, 495–
523.
Taylor, M. S., Masterson, S. S., Renard, M. K., & Tracy, K. B. (1998).
Managers’ reactions to procedurally just performance management
systems. Academy of Management Journal, 41, 568–579.
Tenbrunsel, A. E., & Messick, D. M. (1999). Sanctioning systems, decision
frameworks, and cooperation. Administrative Science Quarterly, 44,
684–707.
Tenenhaus, M., Vinzi, V. E., Chatelin, Y., & Lauro, C. (2005). PLS path
modeling. Computational Statistics & Data Analysis, 48, 159–205.
736 F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737
Tomkins, C. (2001). Interdependencies, trust and information in
relationships, alliances and networks. Accounting, Organizations and
Society, 26, 161–191.
Vandenbosch, M. B. (1996). Con?rmatory compositional approaches to
the development of product spaces. European Journal of Marketing, 30,
23–46.
Whitener, E. M., Brodt, S. E., Korsgaard, M. A., & Werner, J. M. (1998).
Managers as initiators of trust: An exchange relationship framework
for understanding managerial trustworthy behavior. Academy of
Management Review, 23, 513–530.
Withey, M., Daft, R. L., & William, H. C. (1983). Measures of Perrow’s work
unit technology: An empirical assessment and a new scale. The
Academy of Management Journal, 26(1), 45–63.
Williamson, O. E. (1993). Calculativeness, trust, and economic
organization. The Journal of Law and Economics, 36, 453–486.
Wold, H. (1982). Soft modeling – The basic design and some extensions. In
K. Jöreskog & H. Wold (Eds.), Systems under indirect observation II
(pp. 1–53). Amsterdam: North-Holland Press.
Wold, H. (1985). Partial least squares. In S. Kotz & N. L. Johnson (Eds.).
Encyclopedia of statistical sciences (vol. 6, pp. 581–591). New York:
Wiley.
Zaheer, A., McEvily, B., & Perrone, V. (1998). Does trust matter? Exploring
the effects of interorganizational and interpersonal trust on
performance. Organization Science, 9, 141–159.
Zimmerman, J. L. (2003). Accounting for decision making and control.
Boston: Irwin, McGraw-Hill.
F. Hartmann, S. Slapnic?ar / Accounting, Organizations and Society 34 (2009) 722–737 737
doc_323975668.pdf