Professional Documents
Culture Documents
L09 - Merit Pay HRM 1998
L09 - Merit Pay HRM 1998
L09 - Merit Pay HRM 1998
Theories of motivation and common expectation argue that people who contribute more to an
undertaking should receive more from it. This expectation has significantly influenced the de-
sign of compensation systems. Merit pay reflects this influence. Nonetheless, in spite of intuitive
appeal and apparent theoretical support, merit pay rarely achieves its objectives. This article
reviews five common implementation issues. It also argues that merit plans underemphasize
important attribution biases that affect performance judgments. These biases suggest that such
plans would still have limited effectiveness, even if implementation problems were overcome.
The article presents an alternative focused on group accomplishment. © 1998 John Wiley &
Sons, Inc.
Human Resource Management, Summer 1998, Vol. 37, No. 2, Pp. 131–146
© 1998 John Wiley & Sons, Inc. CCC 0090-4848/98/02131-16
132 • H UMAN RESOURCE MANAGEMENT, Summer 1998
mance (Gomez-Mejia & Balkin, 1992, validity and accuracy of the most commonly
Hughes, 1986; Kanter, 1987; Whyte, 1955). used method, supervisory ratings (Milkovich
Although the potential problems with & Boudreau, 1991), has always been a major
merit pay have been widely discussed (e.g., concern. These measures are likely to suffer
Dessler, 1994; Fisher, Schoenfeldt, & Shaw, from both deficiency and contamination prob-
1993; Henderson, 1989), these difficulties lems, including subjectivity, personal bias,
have not lessened its attractiveness. Firms deliberate distortion, and various other inten-
seem to prefer a merit pay approach to other tional and unintentional rating errors (e.g.,
alternatives, and HR professionals have re- halo, contrast, central tendency, etc.). All these
sponded by identifying the conditions “favor- problems call into question the accuracy of
able” to merit pay. These efforts imply that, in the performance measure (Balkin & Gomez-
the right circumstances, merit pay can lead to Mejia, 1987; Latham & Wexley, 1981;
overall increases in motivation and perfor- Longenecker, Sims, & Gioia, 1987).
mance. We believe that, even in favorable cir- The second problem is the issue of “inter-
cumstances, merit pay is unlikely to bring related performances” where the work tasks
about these outcomes. Consequently, firms of two or more individuals are tied together.
may want to explore alternatives to merit pay. Here, the firm typically has great difficulty in
The purpose of this article is threefold. separating the unique contributions of the
First, we examine the problems associated with various individuals. Even when such determi-
merit pay and evaluate proposed solutions. nations appear possible, some investigators
Second, we consider important attribution (Deming, 1986) have questioned whether
errors that attenuate merit pay effectiveness, separation is justified, since individuals were
arguing that these problems have been not entirely in control of their own results.
underemphasized. Finally, we propose a work-
unit based alternative to individually-oriented Traditional Solution. To overcome the first
merit pay systems. difficulty, human resource (HR) researchers
have suggested a dual strategy. First, they have
recommended that firms move away from
Problems Associated with Merit Pay highly subjective judgments to more specific
and objective measurements. Second, they
Several kinds of problems hamper the effec- have suggested that firms train appraisers to
tiveness of merit pay. Most reflect some type avoid the common errors that occur in per-
of implementation hurdle. Others entail po- formance measurement.
tential problems should the firm actually The current shift from measuring individual
implement a merit pay approach. We have clas- traits (e.g., initiative), to measuring job behav-
sified the various difficulties into five catego- iors and job results (Fisher et al., 1993) reflects
ries: (1) measurement issues, (2) performance the first aspect of this dual strategy. The inher-
appraisal (PA) feedback and acceptance issues, ent objectivity of measuring performance against
(3) reward desirability issues, (4) system either output or goals is evident. Where output
“noise” issues, and (5) unintended conse- measures are less appropriate, an array of be-
quences. Of course, a specific merit pay prob- havioral measures is available. These measures
lem may have roots in several categories. include behaviorally anchored rating scales
(BARS); behavioral observation scales (BOS);
and behavioral discrimination scales (BDS).
Measurement Issues These measurement approaches have specifi-
cally evolved to overcome the difficulties asso-
Merit pay measurement issues take two forms. ciated with the graphic rating scale (Carroll &
First, can the firm actually measure individual Schneier, 1982).
performance? Second, can the firm clearly Researchers have also championed rater
distinguish between individuals whose jobs are training as highly effective in reducing mea-
interrelated? For the first question, the answer surement rating errors. Simply by training
is surprisingly equivocal. While firms have raters in how to avoid common errors, firms
many methods for gauging performance, the can make the measurement process more
Merit Pay, Performance Appraisal, and Individual Motivation • 133
accurate and objective (Latham & Wexley, Supervisory Feedback Skills. Giving evalu-
1981); thus, training and a results- or behav- ative feedback is often anxiety-provoking for
iorally-oriented appraisal focus may minimize the appraiser, and managers are often uncom-
measurement problems. fortable with the whole feedback process. Sev-
For interrelated performance, research has eral factors contribute to this. First, the typical
not provided any solution. The literature (e.g., feedback session is characterized by disagree-
Gomez-Mejia & Balkin, 1992) simply suggests ment, making the usual session unpleasant.
that merit pay is a less feasible option under Second, the supervisor is trying to accomplish
these circumstances. Merit pay is limited to several distinct objectives by providing feed-
situations in which managers can make fairly back: that is, evaluating and rewarding past Finally, a
clear performance attributions and employees performance while attempting to develop and considerable
generally perform independent work. motivate future performance. This dual focus amount of skill
makes it difficult to accomplish either of the is needed to
conduct an
objectives well. effective
PA Feedback and Acceptance Issues Finally, a considerable amount of skill is appraisal
needed to conduct an effective appraisal in- interview, and
For merit pay plans to work, employees must terview, and many managers apparently do not many managers
agree that pay distinctions based on the perfor- possess the requisite skill. It is, therefore, not apparently do
not possess the
mance measures are fair. In other words, it is surprising that many managers see little value requisite skill.
not sufficient for the measures simply to be ac- in PA interviews (Napier & Latham, 1986), or
curate; employees must perceive them as accu- that many employees feel that appraisal feed-
rate. For many firms, the problems discussed back causes more confusion than enlighten-
have sensitized employees to regard PA with ment (Meyer, 1991).
suspicion and cynicism—creating two concerns.
Traditional Solution. To increase employee
Employee Acceptance. Employees generally acceptance of PAs (and thus the merit deci-
do not agree with their performance appraisals sions based on them), researchers have rec-
(Campbell & Lee, 1988; Meyer, 1980) and pre- ommended that appraisers involve employees
sumably do not accept the evaluations as accu- in the evaluation process. Evidence
rate. Although the use of observable, objective (DeGregorio & Fisher, 1988) suggests that
criteria lessens this problem, it does not elimi- individuals desire such involvement and that
nate it. As Fisher et al. (1993) have noted, fac- participation may increase acceptance of the
tors beyond the individual’s control (e.g., the appraisal results.
quality of a work machine) can affect even ob- Investigators have also suggested strate-
jective performance. This makes it less likely gies for reducing the problems supervisors face
that an individual will accept a disappointing when giving feedback. Textbooks often pro-
evaluation even when no disagreement exists vide guidelines for appraisers to follow (e.g.,
about the outcomes. The numbers are not in “prepare carefully, provide specific examples,”
dispute; the causes for the numbers are. and so forth). These can reduce anxiety by
Additionally, a complex set of informa- structuring an ambiguous task. They may also
tional, cognitive, and affective constraints in- help develop the skills needed to conduct bet-
fluences self evaluation, and these constraints ter appraisals.
make it extremely unlikely that individuals will For the problem of multiple objectives, it
agree with evaluations lower than their own has long been recommended that separate
(Campbell & Lee, 1988). While deliberate interviews be used for evaluation and devel-
distortion may account for some of the dis- opment purposes (Meyer, 1991; Meyer, Kay,
agreement, more fundamental factors (i.e., & French, 1965). By separating the interviews,
different cognitive schemas; defense mecha- supervisors can focus on past performance and
nisms, etc.) appear to be operating. Because merit pay in one session, and cover develop-
of such disagreement, many employees sim- mental issues and future goals later, when
ply do not see merit pay as truly rewarding defensiveness is lower.
performance (Hills et al., 1987; Kanter, 1987; Are these recommendations effective?
Vest, Hills, & Scott, 1989). Certainly participation and involvement are
134 • H UMAN RESOURCE MANAGEMENT, Summer 1998
helpful, and more skillfully conducted inter- budget enough money to make merit pay
views are useful. Nonetheless, at least as early “work,” but clearly it would prefer to spend
as 1965, Meyer and his colleagues dealt with no more than this. A firm might object to an
the same issues and pointed out the same so- arbitrarily large merit differential determined
lutions. A quarter-century later, he has re- primarily by fiat. Alternatively, cost-benefit
ported that little has changed (Meyer, 1991). analyses of merit pay levels might empirically
One survey of American firms, cited by Meyer determine the appropriate differential. Such
(1991), indicated that less than 10% have suc- analyses, however, are usually too impractical
cessful PA programs; another nationwide sur- to generate.
vey (Meyer, 1991) indicated that the PA system Instead, many firms simply determine
is the firm’s most frequently mentioned HR what they are willing to spend on merit in a
concern. Nor are these results limited to just given year. To ensure that a reasonable merit
American firms. In a survey of 74 companies differential exists, these firms then treat merit
across all sectors of the Singaporean economy, as a zero-sum distribution (e.g., Thompson &
a major consultancy firm found that fewer Dalton, 1970), with a higher-than-average pay
than 30% of employees believed that perfor- increase to one employee requiring a lower-
mance appraisal helped them (Straits Times, than-average pay increase to another. This
1995). Such widespread and consistently dis- approach allows managers to create relatively
couraging results question the efficacy of the large merit differentials even with small merit
recommendations themselves. budgets. It does so, however, by implicitly
evaluating an individual’s job performance
against work group peers, rather than against
Reward Desirability Issues absolute achievement. This approach presum-
ably is not what individuals anticipate when
A workable merit pay system requires the firm calling for merit pay.
to budget sufficient funds for meaningful pay
differentials. Managers must attach suitably
large pay differences to suitably large perfor- System “Noise” Issues
mance differences (Wallace & Fay, 1988).
Debate exists whether this typically occurs. Factors only indirectly relevant to merit pay
Lawler (1981) has suggested that at least a can obscure the firm’s efforts to link pay to
3% pay change is needed for individuals to performance. Such factors represent “back-
notice a difference. He later concluded that ground noise” drowning out the merit pay “sig-
the typical pay differential between poor and nal.” Two factors are particularly troublesome.
outstanding performance is too small (cited
in Henderson, 1989). Timing of Merit Pay. Firms typically pro-
Henderson (1989), however, has sug- vide merit adjustments on an annual basis.
gested the typical merit pay differential be- Thus, most employees experience a significant
tween high and low performers is 4%. Over time lapse between job performances that
ten years, this leads to a sizable difference in warrant merit and the actual reception of merit
overall earnings. Whether employees adopt a pay. Motivational theorists (e.g., Skinner,
ten-year perspective when evaluating a merit 1974) have pointed out that such delays sub-
increase is not known. stantially dilute the efficacy of the reward and
also hide the performance-reward connection
Traditional Solution. Some investigators from the individual.
have proposed paying top performers increases
that are four times the increases paid to poor Reward System Inconsistencies. The firm’s
performers (Fisher et al., 1993). With such a pay system does not represent its only reward
differential, employees would certainly see a system. Managers typically have available a
difference; however, while large merit differ- large number of other rewards that they can
entials can eliminate the reward desirability bestow or withhold (Fisher et al., 1993). If
issue, practical constraints usually operate to supervisors use these non-monetary rewards
restrain merit differentials. A firm needs to carelessly, employees receive mixed signals.
Merit Pay, Performance Appraisal, and Individual Motivation • 135
Unless such rewards are also performance- potentially important aspects of such work. A
based, they can weaken perceived merit merit pay approach may thus cause managers
connections. to concentrate on short-term achievements
Similarly, performance is rarely the only (i.e., those “blossoming” in the merit pay pe-
factor that determines merit pay. These other riod) and even to resist attempts to change
factors (e.g., an individual’s position in the reward-linked, but outdated, activities.
salary range, the existence of pay compression,
pay relationships within the work group, “start- Affective Consequences. Some investigators
ing point” differentials) can also hide the per- (e.g., Schwab, 1974) have raised the possibility
formance-pay link (Balkin & Gomez-Mejia, of a performance-satisfaction tradeoff with merit
1987; Gomez-Mejia & Balkin, 1992). pay. Incentive systems may increase an
individual’s performance but also introduce ex-
Traditional Solution. The noise problems tra tension and stress, thus reducing satisfac-
described above represent inherent character- tion. Others (Gomez-Mejia & Balkin, 1992)
istics of organized life. While closer temporal have speculated that individuals may perceive
connections between performance and pay may merit pay as threatening their personal au-
be extremely attractive, the administrative diffi- tonomy and reducing their freedom of action.
culties in bringing these about make significant As a result, they experience anger and dissatis-
change unlikely. Similarly, organizational com- faction. In a related vein, merit pay also may
plexity makes it unlikely that a firm’s numerous reduce an employee’s “intrinsic motivation,” that
reward systems can always be consistent. Of is, the individual’s intrinsic desire to perform
course, an awareness of these problems may help an otherwise interesting and engaging activity
minimize counterproductive results. (Deci & Ryan, 1985). This long-run loss of in-
Additionally, in PA feedback sessions, trinsic interest may more than offset merit pay’s
firms can go to great lengths to demonstrate short-run motivational gains.
and clarify performance-reward connections.
In this regard, noise problems highlight the Traditional Solution. Since the above are
importance of individual trust in the perfor- primarily potential problems, researchers have
mance appraisal system. Without such trust, not examined them in any depth. This is par-
a supervisor will be unlikely to convince indi- ticularly true for affective consequences. For
viduals that contradictions in the merit sys- example, some investigators simply question
tem are just random fluctuations. the applicability of Deci’s work to the typical
work organization (e.g., Staw, 1977). Thus, the
absence of solutions to questionable problems
Unintended Consequences is understandable.
The unintended consequences of single-
Another set of problems may arise after a firm mindedness, inflexibility, and so forth are less
adopts a merit pay approach. Such problems easily dismissed. Researchers have frequently
represent potential costs that can negate the asserted that firms “get what they measure.”
benefits of merit pay. The problems are By implication, firms do not get what they do
grouped into two categories. not measure. Although a firm should measure
Researchers
all truly important activities, in practice, this have frequently
Task-Focus Consequences. Merit pay can expectation is unrealistic. Given the rapidly asserted that
make individuals “single-minded” and inflex- shifting, temporally-specific requirements of firms “get what
ible about performance (Gomez-Mejia & the modern firm, comprehensive measure- they measure.”
Balkin, 1992). Employees may only engage in ment is an ideal toward which the organiza-
those activities directly related to reward at- tion strives, not a goal which it achieves.
tainment and neglect other important but less Consequently, some investigators have
tangible aspects of the job (e.g., properly so- argued that a tight connection between per-
cializing newcomers to the firm). This poten- formance and reward is rarely justified
tial problem is exacerbated in complex (Gomez-Mejia & Balkin, 1992). They have
managerial and professional jobs because no suggested that it is better if the pay-perfor-
measurement system can truly capture all the mance link is only “loosely coupled.” Then,
136 • H UMAN RESOURCE MANAGEMENT, Summer 1998
mission, goals, and objectives. This is analo- Since even a small work unit has numer-
gous to conducting a thorough job analysis but ous products and customers, a single overall
at the work unit level. Typically, the group uses PI is usually inadequate. On the other hand,
a structured communication/decision making an attempt to track all potentially relevant PIs
technique, allowing members to consensually may result in information overload. Some re-
identify the unit’s fundamental mission. searchers have suggested that groups develop
The group captures its mission in a writ- a “family” of PIs (a group of separate but mean-
ten statement and uses it to identify the work ingfully related measures) for the 20% of unit
unit’s goals and “customers.” Members then products and services that account for 80% of
analyze the work unit from its clients’ perspec- the success of the unit. Others (Globerson,
tives. For each of the unit’s customers, the Globerson, & Frampton, 1991) have recom-
group identifies the unit’s products and ser- mended a maximum of seven central PIs.
vices most valued by that customer. These rep- Several investigators provide examples of
resent the unit’s key outputs, those products PIs relevant to particular work groups, and
and services most responsive to meeting cus- Table II illustrates some of these PIs.
tomer expectations. Measures of these prod-
ucts must form part of the unit’s PI system.
The group also reviews the functional pro- Benefits of a Unit-Level PI Approach
cesses of the unit, that is, the unit operations
that transform inputs into outputs. This ex- This section examines the potential advantages
amination exposes the procedures important of a PI approach.
for the transformation process and the key
inputs vital for creating the unit’s products and
services. By the end of these steps, the group Measurement Benefits
has systematically reviewed the purposes,
products, and processes of the unit and has The use of performance indicators addresses the
developed a shared understanding of areas and two broad measurement problems confronting
activities critical for unit performance. At its traditional individual appraisal: subjectivity and
best, this analysis stimulates members to con- interdependency. As noted, PIs are expressed
sider not only what the unit actually does, but as some type of statistical ratio. This heavy em-
also what it should be doing. phasis on objectivity and quantification mirrors
the recommendations of HR specialists and
embodies trends already evident in traditional
The Ratios performance evaluation. By moving away from
subjective judgments and substituting objective
The actual PIs are ratios of outputs to inputs. ratios, many of the errors associated with rating
The unit initially creates an exhaustive list of measures (e.g., leniency, halo, and so forth) are
output measures and then selects the most no longer relevant. Even deliberate distortion is
promising of these potential measures. Selec- less likely because of the statistical nature of
tion is based on (1) the measure’s importance the PIs.
and centrality, (2) the cost and effort required Further, because the firm evaluates at the
to obtain the measure, (3) the measure’s “cycle” group level, judgmental problems associated
time (i.e., the time needed for a change to reg- with individual interdependencies are also ir-
ister on the measure), (4) its sensitivity to small relevant. Since unit performance ultimately
changes, and (5) its relative subjectivity. determines each member’s merit pay, the firm
Similarly, the unit carefully generates, avoids having to sort out degrees of perfor-
evaluates, and selects input measures. Focus mance at the individual level.
is on those inputs associated with the produc- Of course, the use of PIs assumes that
tion of outputs selected for tracking. Common the firm has carefully determined the unit’s
input measures include hours worked, unit important performance dimensions, avoiding
payroll expenditure, raw material used, energy both deficiency errors (i.e., the omission of
consumed, and so forth. an important element that should have been
140 • H UMAN RESOURCE MANAGEMENT, Summer 1998
Finance Department
1. Invoices Processed/Invoicing Hours
2. Invoicing Errors/Invoices Processed
3. Number of Pricing Proposals/Number of Pricing People
4. Incomplete Cost Standard/Total Cost Standards
Production Department
1. Wait Time Hours/Direct Labor Hours
2. Customer Accepted Lots/Lots Submitted
3. Warranty Repair Costs/Sales
4. Hours on Labor Ticket Rejects/Total Hours Reported
Marketing Department
1. Change in Sales/Change in Marketing Expenditures
2. Product Awareness Level/Product Marketing Budget
3. Market Share/Promotional Expenditures
included) and contamination errors (i.e., the defensiveness is greatly reduced. Since mem-
inclusion of an irrelevant element). Addition- bers are jointly responsible for performance,
ally, reliability and validity are as fundamental no individual is singled out for weak achieve-
for good PIs as these characteristics are for tra- ment. Poor unit performance becomes a prob-
Performance ditional individual measures. Finally, similar to lem requiring a solution rather than a threat
Indicators have
a number of individual appraisal, useful PIs are sensitive to to be deflected. Second, PIs provide feedback
advantages in small changes in performance; they are under- without the interpersonal evaluation compo-
terms of feedback stood; and they are responsive to the actions of nent present in traditional appraisal. The work
and acceptance. unit members (Campbell et al., 1995). Although design literature suggests that direct, unfil-
PIs provide measurement benefits that go be- tered information from the task itself is supe-
yond individual measures, they have much in rior to supervisory feedback. PIs supply this
common with such measures. type of feedback. Additionally, the quality of
the feedback is independent of supervisory
communication and interpersonal skills.
Acceptance and Feedback Benefits Because the group itself helped create the
PIs, members are likely to accept results based
PIs have a number of advantages in terms of on such measures as meaningful and relevant.
feedback and acceptance. First, because eval- Additionally, with PIs, work units typically
uation is at the work unit level, individual receive feedback more frequently than is
Merit Pay, Performance Appraisal, and Individual Motivation • 141
feasible with individual appraisal. This gives a broad perspective and to think of achieve-
the firm more opportunity to acknowledge ment in terms of the unit’s mission. With PIs,
good results (Campbell et al., 1995). each individual’s responsibility encompasses
all the unit’s tasks, since members are re-
warded or penalized collectively. Thus, mem-
Reward Benefits bers do whatever is necessary to improve
performance rather than only doing “re-
While PIs cannot create more (or more desir- warded” tasks.
able) rewards, their use may help the firm In terms of affective problems, the par-
avoid problems associated with reward desir- ticipative nature of the PI generation process
ability. Specifically, a PI system allows the firm usually ensures high levels of intrinsic satis-
to closely examine the relationship between faction among unit members. This commonly
budgets and performances for various units. manifests itself in increased identification with
Relative to individual evaluation, such inspec- the group and in enhanced commitment to
tions can be more thoughtful and are more group undertakings. The use of PIs, therefore,
likely to reveal inappropriate levels of merit has the potential to enhance cohesion and in-
allocations. Also, with group level PIs, the dividual interest.
problem of rewarding one high performer at Of course, PIs can generate their own set
the expense of another is pushed to the unit of unintended consequences. Werther, Ruch,
level, where the firm has a better chance of and McClure (1986, p. 339) offer an example
detecting and correcting it. of this in describing a set of possible PIs for
the service staff of a new restaurant. Table III
contains five possible PIs from many that the
Noise Reduction unit might have constructed. If Ratio 5 (Dol-
lar Amount of Food Served/Number of Employ-
Noise in an appraisal system is usually linked ees per Shift) defines unit performance, then
to reward timing and to conflicts among re- staff is likely to encourage customers to order
ward systems. Such problems are particularly both more dishes and more expensive dishes.
likely with traditional appraisal because of the If Ratio 4 (Number of Checks/Number of Em-
sheer number of individuals involved. With ployees per Shift) defines performance, then
unit-level PIs, the firm monitors far fewer units staff is likely to find ways to write separate
and has a better chance of minimizing these checks for each individual at a table or for each
difficulties. Also, with PIs, firms can informally item ordered. If Ratio 1 (Number of Custom-
reward superior performance whenever unit ers Served/Number of Employees per Shift)
performance truly warrants such recognition. defines performance, staff will find ways to
By making timely reinforcement feasible, the discourage customers from lingering. Similar
PI system can keep the performance-reward kinds of implications are associated with the
link salient in the minds of employees, even if other PIs.
the formal merit adjustments only occur on Thus, although the PIs appear similar,
an annual basis. their implications are different. A careless-
The use of PIs can also dampen the likeli- ly constructed PI can have a significant
hood of serious reward inconsistencies. Be- negative impact on the unit’s mission. Even
cause the firm links the primary reward system appropriate PIs can have inappropriate con-
to group level accomplishment, reward system sequences. For example, Ratio 1 was not
conflicts occurring at the individual level are meant to hurry customers but to ensure
usually less conspicuous and damaging. prompt and speedy service. As noted earlier,
the use of a family of PIs (in this case, per-
haps Ratio 1 and Ratio 5) can minimize such
Reduction of Unintended Consequences unintended consequences.
Finally, the PIs in Table III are all only
Individual merit pay can encourage employ- “partial” measures. Each is a ratio of output
ees to focus only on tasks that are rewarded. to just one of four categories of inputs (labor,
In contrast, PIs encourage employees to adopt material, cost, or energy). Partial measures can
142 • H UMAN RESOURCE MANAGEMENT, Summer 1998
fail to detect important trade-offs. If the unit These important criticisms are examined
is purchasing speedy service at the cost of in- below.
creased spoilage of material, this trade-off may
not be detected even by a family of PIs such
as those in Table III. Because those PIs are Non-Representative Individual Performance
all partial measures focused only on labor, the
group needs to include one or more PIs fo- The norm of distributive justice argues that an
cused on material as input. Table IV contains individual should be rewarded in proportion to
such a family of partial measures developed his/her contribution. In many situations, the use
for a sales unit. of unit PIs will violate the norm. Employees who
are not representative of their unit (both stars
and social loafers) will receive more or less than
Discussion their individual efforts truly warrant.
This raises a question as to whether the
Although the use of PIs appears to alleviate elementary unit of the firm is the individual
many merit pay difficulties, an analyst might or the work group. Although noted theorists
raise at least two concerns. First, the approach (e.g., Likert, 1967) have periodically champi-
ignores individual performance. If merit oned the work unit, firms hire individuals, not
pay’s purpose is to motivate the individual, the groups, and individuals see themselves as con-
suggested scheme may discourage outstand- tributors distinct from the work unit. None-
ing individual performers since their efforts theless, the success of the firm rests squarely
go unrewarded in an average performing on group performance. In this light, the firm’s
unit. traditional concern with recognizing and re-
Second, group evaluation may solve the warding individual merit is misplaced. If the
problems discussed earlier only by elevating work unit is at the core of achievement, then
those problems to the work unit level. If so, the firm’s central reward and recognition ef-
the firm still has to grapple with the issues. forts must also be at this level.
For stars, this orientation requires a acceptable PIs have been created. At this
change in how they view personal accomplish- point, what then happens? Does the firm ease
ment. Under the old system, accomplishment into the new system or does it completely aban-
merely entails being good at one’s task. With don its individual performance system for the
PIs, it now also entails being good at helping new approach?
others be good at their tasks. By creating this Easing into the new procedures (e.g., by
mind-set among employees, the apparent vio- augmenting the existing system with the recently
lation of distributive justice is much reduced. created PIs) has several benefits. The company
At the other extreme, work unit dynamics minimizes disruption and gets a chance to test
often develop that prevent distributive justice the new PIs. Further, some experts (e.g.,
violations through social loafing. In addition Heneman, 1992) have argued that using such
to experiencing monetary and social pressure hybrid systems is highly appropriate. Since dif-
to perform acceptably, individuals often en- ferent performance management systems serve
counter an element of ego. Unit members can different purposes (e.g., one motivating groups;
typically identify individuals whose perfor- another, individuals), a firm may find it useful
mance is continuously below par, and few in- to have multiple systems in place. For example,
dividuals want their cohorts to view them as Pacific Gas and Electric (1989) has performance
incompetent. systems that reward corporate, work group, and
individual achievement; and, similarly, Lincoln
Electric manages performance through a com-
Work Unit Level Problems bination of individual and group level evalua-
tions (Lienert, 1995; Perry, 1990; see also Shaw
The second concern is that PIs create regres- & Schneier, 1995).
sion problems: Issues resolved at the individual Nonetheless, a firm must carefully exam-
level reappear at the group level. Instead of ine this intuitively appealing notion. Conflicts
problems centered on individual task interde- among different reward systems often cause
pendencies, the firm now faces problems fo- problems with traditional merit approaches, and
cused on group interdependencies; or, instead these difficulties may hamper a formal hybrid
of individual complaints about reward ineq- system. Multiple systems, with their different
uities, criticisms now arise from work units. emphases, can create confusion and conflict
These concerns are partially justified. In without extremely careful management.
solving one set of difficulties PI systems may Companies may find an alternative strat-
create others. The firm still benefits, however, egy useful. Management must first fully ex-
because problems at the group level are likely amine the implications of a PI system to
to be fewer in total (given the smaller number ensure its compatiblity with both organiza-
of units involved) and more noticeable. Since tional culture and employee expectations.
a specific problem necessarily affects a larger Then, having developed PIs, units should track
organizational unit, it is less easily missed or performance for several quarters, ensuring
ignored (cf., Brown, 1984; Gomez-Mejia & that the system functions as expected. Finally, As noted earlier,
group members
Balkin, 1992). Thus, the firm has a better the work groups should switch completely to typically are
chance of detecting and correcting merit is- the PI system on a date they previously deter- heavily involved
sues at this level. mined. This kind of transitional approach, al- in the
lowing the groups to determine the development of a
change-over period, makes use of the team- Performance
Indicators (PI)
Transitioning to Group Performance building potential inherent in a PI system, and system, so
Management also avoids the problems associated with mul- transitional
tiple performance management systems. processes begin
In moving to PIs, firms must carefully con- with this stage
sider the transition from one system to an- and continue
until acceptable
other. As noted earlier, group members Conclusion
PIs have been
typically are heavily involved in the develop- created.
ment of a PI system, so transitional processes To remain competitive, a firm needs to entice
begin with this stage and continue until employees to perform at their best. While
144 • H UMAN RESOURCE MANAGEMENT, Summer 1998
merit pay plans appear to represent one pow- on PIs, while not as intuitively obvious, po-
erful and intuitively appealing enticement, for tentially has a great deal to offer. Companies
a substantial number of firms using such dissatisfied with the traditional merit approach
plans, the reality is much different from the may wish to explore this alternative.
appearance. A unit-level merit system based
HO-BENG CHIA is a lecturer at the National University of Singapore and teaches organi-
zational behavior, human resource management, creative problem solving, and negotia-
tions and mediation. He received his doctorate in organizational behavior from the Univer-
sity of British Columbia. His research interests include goal setting and motivation, conflict
management, cross-cultural differences in organizational behavior, and organizational
effectiveness.
REFERENCES
Anderson, J., & Camealy, J. (1991). Corporate opera- Campbell, D., Campbell, K., & Chia, H.B. (1995).
tional analysis. Westport, CT: Quorum Books. Innovation through performance analysis:
Balkin, D., & Gomez-Mejia, L. (1987). New perspec- Teambuilding and the development of perfor-
tive on compensation. Englewood Cliffs, NJ: mance indicators. Innovation and Creativity Man-
Prentice-Hall. agement, 4, 160–171.
Bernardin, H.J., & Beatty, R. (1984). Performance Campbell, D., & Lee, C. (1988). Self appraisal in
appraisal: Assessing human behaviour at work. performance evaluation: Development versus
Boston, MA: Kent Publishing Co. evaluation. Academy of Management Review, 13,
Bierhoff, H. (1989). Person perception and attribu- 302–314.
tion. Berlin: Springer-Verlag. Carroll, S., & Schneier, C. (1982). Performance Ap-
Brinkerhoff, R., & Dressler, D. (1990). Productivity praisal and Review Systems. Glenview, IL: Scott,
measurement: A guide for managers and evalua- Foresman & Co.
tors. Newbury Park, CA: Sage Publications. Deci, E., & Ryan, R. (1985). Intrinsic motivation and
Brown, K. (1984). Explaining group performance: An self-determination in human behavior. New York:
attributional analysis. Academy of Management Plenum Press.
Review, 9, 54–63. DeGregorio, M., & Fisher, C. (1988). Providing per-
Brown, K., & Mitchell, T. (1986). Influence of task formance feedback: Reactions to alternate meth-
interdependence and number of poor perform- ods. Journal of Management, 14, 605–616.
ers on diagnoses of causes of poor performance. Deming, W.E. (1986). Out of crisis. Cambridge, MA:
Academy of Management Journal, 29, 412–424. Productivity Press.
Merit Pay, Performance Appraisal, and Individual Motivation • 145
Dessler, G. (1994). Human resource management. Knowlton, W., & Mitchell, T. (1980). Effects of causal
Englewood Cliffs, NJ: Prentice-Hall. attributions on a supervisor’s evaluation of sub-
Edmunds, S. (1982). Performance measures for grow- ordinate performance. Journal of Applied Psychol-
ing businesses. New York: Van Nostrand Reinhold ogy, 65, 459–466.
Co. Latham, G., & Wexley, K. (1981). Increasing produc-
Fisher, C., Schoenfeldt, L., & Shaw, J. (1993). Hu- tivity through performance appraisal. Reading,
man resource management. Boston, MA: MA: Addison-Wesley.
Houghton Mifflin. Lawler, E. (1981). Pay and organizational development.
Fiske, S., & Taylor, S. (1984). Social cognition. Read- Reading, MA: Addison-Wesley.
ing, MA: Addison-Wesley. Lienert, A. (1995). A dinosaur of a different color.
Gioia, D. (1989). Self-serving bias as a self- Management Review, 84(2), 24–29.
sensemaking strategy: Explicit versus tacit im- Likert, R. (1967). The Human Organization. New York:
pression management. In R. Giacalone & P. McGraw-Hill.
Rosenfeld (Eds.), Impression Management in the Longenecker, C., Sims, H., & Gioia, D. (1987). Be-
Organization (pp. 219–234). Hillsdale, NJ: hind the mask: The politics of employee appraisal.
Erlbaum. Academy of Management Executive, 1, 183–191.
Globerson, A., Globerson, S., & Frampton, J. (1991). Meyer, H. (1991). A solution to the performance ap-
You can’t manage what you don’t measure. praisal feedback enigma. Academy of Management
Aldershot, England: Avebury Press, Gower Pub- Executive, 5, 68–76.
lishing Co. Meyer, H. (1980). Self appraisal of job performance.
Gomez-Mejia, L., & Balkin, D. (1992). Compensa- Personnel Psychology, 33, 291–295.
tion, organizational strategy, and firm performance. Meyer, H., Kay, E., & French, J. (1965). Split roles in
Cincinnati, OH: South-Western Pub. Co. performance appraisal. Harvard Business Review,
Heider, F. (1958). The psychology of interpersonal re- 43, 123–129.
lations. New York: John Wiley & Sons. Milkovich, G., & Boudreau, J. (1991). Human resource
Henderson, R. (1989). Compensation management: management. Homewood, IL: Irwin.
Rewarding performance. Englewood Cliffs, NJ: Mitchell, T., Green, S., & Wood, R. (1981). An
Prentice-Hall. attributional model of leadership and the poor
Heneman, R. (1992). Merit pay. Reading, MA: performing subordinate: Development and vali-
Addison-Wesley. dation. In B. Staw & L. Cummings (Eds.), Re-
Heneman, R., Greenberger, D., & Anonyuo, C. search in Organizational Behavior, Vol. 3,
(1989). Attributions and exchanges: The effects 197–234. Greenwich, CT: JAI Press.
of interpersonal factors on the diagnosis of em- Mullen, B., & Riordan, C. (1988). Self-serving attri-
ployee performance. Academy of Management butions for performance in naturalistic settings:
Journal, 32, 466–476. A meta-analytic review. Journal of Applied Social
Hills, F., Scott, D., Markham, S., & Vest, M. (1987). Psychology, 18, 3–22.
Merit pay: Just or unjust desserts. Personnel Ad- Napier, N., & Latham, G. (1986). Outcome expect-
ministrator, 32, 53–64. ancies of people who conduct performance ap-
Hughes, C. (1986). The demerit of merit. Personnel praisals. Personnel Psychology, 39, 827–837.
Administrator, 31, 40. Pacific Gas and Electric Co. (1989). The 1989 Man-
Ilgen, D., Mitchell, T., & Fredrickson, J. (1981). Poor agement Incentive Program and Information on
performers: Supervisors’ and subordinates’ re- the 1989 Merit Plan and Performance Recogni-
sponses. Organizational Behavior and Human tion Program. San Francisco, CA: Author (cited
Performance, 27, 386–410. in Heneman, 1992).
Johns, G. (1994). Absenteeism estimates by employ- Perry, T. (1990). Staying within the basics. HR Maga-
ees and managers: Divergent perspectives and zine, 35(11), 73–76.
self-serving perceptions. Journal of Applied Psy- Porac, J., Ferris, G., & Fedor, D. (1983). Causal attri-
chology, 79, 229–239. butions, affect, and expectations for a day’s work
Jones, E., & Nisbett, R. (1972). The actor and ob- performance. Academy of Management Journal,
server: Divergent perceptions of causes of behav- 26, 285–296.
ior. In E. Jones et al. (Eds.) Attribution: Perceiving Porter, L., & Lawler, E. (1968). Managerial attitudes
the Causes of Behavior. Morristown, NJ: General and performance. Homewood, IL: Dorsey-Irwin.
Learning Press. Ross, J. (1993). Total quality management. Delray
Jowett, P., & Rothwell, M. (1988). Performance indi- Beach, FL: St. Lucie Press.
cators in the public sector. London: MacMillan Sanna, L., & Mark, M. (1995). Self-handicapping,
Press. expected evaluation, and performance: Accentu-
Kanter, R. (1987). From status to contribution: Some ating the positive and attenuating the negative.
organizational implications of the changing ba- Organizational Behavior and Human Decision
sis for pay. Personnel, 64(1), 12–37. Processes, 64, 84–102.
146 • H UMAN RESOURCE MANAGEMENT, Summer 1998
Schwab, D. (1974). Conflicting impacts of pay on Behavior (pp. 55–95). Chicago, IL: St. Clair.
employee motivation and satisfaction. Personnel Straits Times. (1995). Appraisal systems: How effec-
Journal, 53(3), 190–206. tive are they?, 10th Feb., 10.
Shaw, D., & Schneier, C. (1995). Team measurement Thompson, P., & Dalton, G. (1970). Performance
and rewards: How some companies are getting it appraisal: Managers beware. Harvard Business
right. Human Resource Planning, 18(3), 34–49. Review, Jan.–Feb., 149–157.
Shimizu, M., Wainai, K., & Nagai, K. (1991). Value Vest, M., Hills, F., & Scott, K. (1989). Determinants
added productivity measurement and practical of instrumentality beliefs in a merit pay system.
approach to management improvement. Tokyo: Paper presented at the Academy of Mgmt Meet-
Asian Productivity Organization. ing, Washington, DC.
Sink, D.S. (1985). Productivity management: Plan- Vroom, V. (1964). Work and motivation. New York:
ning, measurement and evaluation, control and John Wiley & Sons.
improvement. New York: John Wiley & Sons. Wallace, M., & Fay, C. (1988). Compensation theory
Sink, D.S., & Tuttle, T. (1989). Planning and mea- and practice. Boston: PWS-Kent Publishing Co.
surement in your organization of the future. Watson, D. (1982). The actor and the observer: How
Norcross, GA: Industrial Engineering and Man- are their perceptions of causality divergent? Psy-
agement Press. chological Bulletin, 92, 682–700.
Skinner, B.F. (1974). About behaviorism. New York: Werther, W., Ruch, W., & McClure, L. (1986). Pro-
Knopf. ductivity through people. St. Paul, MN: West
Staw, B. (1977). Motivation in organizations: Toward Publishing Co.
synthesis and redirection. In B. Staw & G. Whyte, W. (1955). Money and motivation. New York:
Salancik (Eds.), New Directions in Organizational Harper & Row.