International Review of Public Administration 29

2007, Vol. 11, No. 2

DEVELOPMENT OF A DIAGNOSIS MODEL FOR PAY-FOR-PERFORMANCE

IN THE PUBLIC SECTOR AND ITS APPLICATION:
FOCUSED ON AGENCIES*

DO LIM CHOI
Chungnam National University, South Korea

This study develops a model to diagnose and examine current organizational status in
preparation for the introduction of a pay-for-performance system that can be applied to reality.
The findings are as follows. First, employee participation, rewards, measurability, and
performance appraisal are core factors. Second, employee participation and measurability
are prerequisites or fundamental factors. Third, performance appraisal and rewards are
dependent on these prerequisites or fundamental factors. Fourth, and more specifically, the
degree of employee participation and measurability directly influence the effectiveness of
rewards and performance appraisal. From a comprehensive perspective, it is necessary for
agencies in Korea to make more effort and preparation to achieve successful pay-for-perfor-
mance systems.

Key Words: Pay-for-Performance, Performance Management, Motivation

INTRODUCTION Motivating employees is among these ten tasks, and to

achieve improved employee motivation, many coun-
Only people who received it originally are paid it again. tries, including the United States, Great Britain, Aus-
Everybody already knows the result of the own appraisal tralia, and New Zealand, have introduced pay-for-per-
(Choi, 2004). (Remarks about pay-for-performance heard formance systems along with civil service reforms.
from a public administrator) Through introducing these systems and the associated
values within the public sector, organizations intended to
During the last few decades, performance measure- motivate employees to act innovatively and sponta-
ment has become a concern in all organizations. Every neously to achieve organizational goals. The current
area of organization management, particularly those assessment is that the success of these systems in public
areas concerned with budgeting, strategic management, sector agencies is dubious at best (Ha, 2004; Young,
and motivation, pays attention to measuring individual 2003; Perry, 1995). Scholars point out that the further
and organizational performance. The introduction of pri- introduction of pay-for-performance systems is not
vate sector management techniques has drawn, and will going to be easy (Risher, 2004; Dakahashi, 2000;
continue to draw, a great deal of attention to perfor- Pearce, Stevenson, and Perry, 1985; Deci, 1975; Meyer,
mance measurement (Herbert, 2004). 1975).
There are many reasons to measure individual and As one can infer from the words of the administrator
organizational performance. Hatry (1999) suggests that quoted above, some problems that have resulted from
public sector managers can use information from perfor- pay-for-performance systems have been discussed by
mance measurement to perform ten different tasks. scholars interested in this area. It is not easy to find a
30 Development of a Diagnosis Model for Pay-for-Performance in the Public Sector and Its Application Vol. 11, No. 2

systematically organized model to diagnose and analyze ability of performance appraisal (Kang & Kwon, 2004),
the status of an organization that has already made the culture or experience such as seniority (Hwang, 2003;
decision to introduce pay-for-performance systems and Hwang, 2003; Young, 2003) the ability and qualifica-
now wishes to implement them. tions of supervisors (Ingraham, 2005; Young, 2003), the
The purpose of this study is to develop a model to motivation of supervisor (Ingraham, 2005; Young,
diagnose and examine current organizational status with 2003), work flow (Young, 2003; Roberts, 1997), and the
respect to the introduction of pay-for-performance and type and amount of reward (Ingraham, 2005; Young,
to apply the model to the reality using two agencies as 2003).
objects of applying the diagnosis model in the practical Based on theories or previous research, several core
arena. variables can be elicited that influence the success and
satisfaction of pay-for-performance. These core vari-
ables, used to build a diagnosis model for pay-for-per-
THEORETICAL BACKGROUND formance in a later section will cover all relevant influ-
ential factors.
Theoretical Origins of Pay-for-Performance The first core variable is a meaningful reward. Finan-
cial rewards given after performance appraisal should be
Pay-for-performance means that employees achiev- meaningful to public servants. Public servants ought to
ing higher performance receive greater rewards. Merit- be motivated to work hard because of the meaningful-
pay1 programs originate from widely accepted theories ness of reward type and amount. The type of reward is
of individual motivation, including Vrooms (1964) already determined and the amount of reward will then
expectancy theory and Adams (1963) equity theory. matter. On the other hand some insist that pay can be a
The expectancy theory assumes that individuals who critical support but not a main driver. Pay is just one of
expect to receive a valued reward for high performance several issues that must be addressed during any major
are more likely to strive for that level of performance change initiative (Flannery, Hofrichter, & Platten, 1996).
than individuals who do not expect to receive any incen- The second variable is fair performance appraisal.
tives. This implies the following. First, financial rewards Fair appraisal for performance will be a prerequisite, in
are one of the most useful tools to motivate employees order to provide fair rewards with which public adminis-
in organizations. Second, through implementing pay- trators are satisfied. Distorted or unfair performance
for-performance, organizations or leaders can expect appraisal can occur from two aspects. The first aspect is
that employees will invest more effort in work and pro- culture or experience. It is assumed that strong patron-
duce higher productivity than in the past. Third, in order age controls organizational decision-making. It will be
for pay-for-performance programs to be effective, hard in that case to expect to achieve a fair appraisal.
employees should place a high value on financial The second aspect is the readiness for fair appraisal.
rewards (OECD, 1993). Although the organizational culture is fair enough and
Equity theory, on the other hand, posits that individu- appraisers have enough qualifications, the work perfor-
als weigh what they put into a job situation (input) mance of public servants ought to be able to be com-
against what they get from it (outcome) and then com- pared quantitatively. Otherwise, there will be high possi-
pare the input-outcome ratio with the input-outcome bility that employees will resist the results of perfor-
ratio of relevant others. If they perceive the ratio to be mance appraisal. In addition the traditional appraisal
equal to that of the relevant others with whom they com- method which is rather one-dimensional must be
pare themselves, a state of equity is said to exist. If the replaced by more comprehensive and dynamic measure-
ratios are unequal, inequity exists. In the state of equity, ment tools that will provide an accurate readout of
employees will be more satisfied and will work hard. employee performance. Performance management
should be aligned with and support the organizational
Critical Factors for the Success culture (Flannery, Hofrichter, & Platten, 1996).
of Pay-for-Performance The third core variable is employee participation in
the process of decision-making related to pay-for-perfor-
There are many factors discussed for the success of mance. As in other organizational activities employee
pay-for-performance, among them - fairness and accept- participation is a critical factor to motivate public ser-
January 2007 Do Lim Choi 31

vants through merit-pay systems. To provide valuable Previous Research on Pay-for-Performance

rewards to public administrators it is important to know
what and how much is wanted. Participation is a good Previous studies are briefly reviewed in this section
tool to get the information. Flannery et al. (1996) sug- to identify existing research issues on pay-for-perfor-
gest nine principles of dynamic pay and one of them is mance and to show justification for this study. As can be
to democratize the pay process. They insist that many of seen in Table 1, articles from five major journals dealing
the ongoing day-to-day activities surrounding compen- with public administration in Korea have been reviewed
sation should be handled at the line, the department, or from 1998. Several significant factors can be identified
the team level. from the literature review.
The final critical variable is measurability. To deter- It is difficult to find studies that examine the relation-
mine the level of merit-pay, work performance must be ships among critical variables relevant to pay-for-perfor-
measured. There are many kinds of performance mea- mance system, as most quantitative studies use frequen-
sures. For example, Hatry (1999) suggests cost mea- cy and a means analyses in order to measure the level of
sures, workload-accomplished measures, effectiveness employee satisfaction. Second, it is not easy to find stud-
measures, and efficiency measures. One function of per- ies that offer a model to diagnose and assess pay-for-
formance measurement is appraising.2 Generally speak- performance systems in any organization. Therefore, a
ing, the high possibility of quantitative measurement is systematic approach for pay-for-performance is difficult,
demanded to objectively appraise the performance of and organizations spend valuable time attempting to
public servants. One characteristic of government activi- answer the following question: what are the prerequi-
ties is that it is difficult to transfer them to market value sites for the success of a pay-for-performance system
(Bruijn, 2002). and what issues should be addressed later?

Table 1. Review of Previous Studies of Pay-for-Performance

Author Variables researched Main contents Approaches

Competition and cooperation as
Lee, H. W. Organizational commitment, trust,
factors to influence the improvement Behavioral, Descriptive
(2005) information sharing, and merit pay
of organizational efficiency
Lee, H. T. Acceptability, rationality of selection Comprehensive and empirical analyses Behavioral, Quantitative
(2005) process, valence, effectiveness and political implications of merit pay (frequency and mean analyses)
Diagnoses of performance management Behavioral, Quantitative
Ha, M. S. Values, necessities, intentions,
system in government and finding (frequency, cross-tab, mean,
(2004) purposes, and effectiveness.
alternatives and priority analyses)
Kang, H. S. & Adequacy, validity, incentive attraction, Improvement of acceptability of the
Behavioral, Quantitative
Kwon, Y. S. system understanding, attitude, and results of performance evaluation
(structural equation model)
(2004) acceptability in governments
Hwang, S. W. Equity, measurement error, and Finding problems in merit-pay system
Systematic, Descriptive
(2003) heterogeneity and improvement
Attraction, expectancy, fairness,
Oh, S. H. Finding problems in merit-pay system Behavioral, Quantitative
self-evaluation, interview, group
(2003) and improvement (frequency)
incentive, evaluation training, IT
Individualism, collectivism,
Behavioral, Quantitative
Hwang, C. Y. authoritarianism, democracy, Relationship between organizational
(factor, variance, correlation,
(2003) performance performance appraisal, culture and performance management
LSD analyses)
merit-pay, and promotion
32 Development of a Diagnosis Model for Pay-for-Performance in the Public Sector and Its Application Vol. 11, No. 2

Pay-for-Performance in Agencies addition, many state and local governments reported

many obstacles in the process of implementing merit-
In this study survey respondents were sampled from pay systems. These disappointments notwithstanding,
two agencies which are part of the agency systems in performance pay schemes have spread rapidly in recent
Korea. In 2000, the agency system was introduced into years. Many OECD countries, including Korea, have
the Korean government. The economic crisis allowed introduced merit pay systems as a means to change the
the IMF (International Monetary Fund) to become stagnating atmosphere of the public sector employment
involved in the internal affairs of Korea and the Korean environment. In this context, in order to achieve fruitful
government introduced the agency system to meet the results from pay-for-performance type models that can
changing economic environment. diagnose the current motivational and performance mea-
The agency system was adopted for government suring situations of the organizations will be suggested
work that fits the market-oriented competition principle based on previous studies.
of maintaining publicness. Agency chiefs were given
administrative and financial autonomy or discretion and Vertical Dimension:
took responsibility for the results of the operations. This Participation and Meaningful Rewards
was a turning point for transforming governmental orga-
nizations into a performance-oriented system. With the introduction of new systems, a high degree
The main reason agencies were used as a sample is of employee participation in system design and adminis-
that the work characteristics are the most market-oriented tration will be an important factor. It will facilitate pat-
ones in the public sector and reflect the operations of pri- terns of interaction that are conducive to the success of
vate companies. There was a high possibility that pay- pay-for-performance plans. In other words, participation
for-performance would be more effectively implemented will help employees accept and understand new changes
in this sphere than in other governmental organizations. occurring in the organizations. Howard Risher (2004)
The main characteristics of agencies are: responsibil- argues that introduction of a pay-for-performance sys-
ity, freedom, and competition (You, 1999; Kim, 1999). tem is likely to be successful if managers and employees
Compared to other governmental organizations, they understand why the change is necessary, what they can
have more discretion in personnel management. For expect, and what steps are planned to ensure it is fair
example, they can offer public servants special pay-for- and equitable. Without employee participation, top-
performance and promotions according to the discretion. down decision-making about merit-pay systems will
The central government shows a tendency to provide prevent managers and employees from believing that
agencies with more autonomy in the operation of the they have ownership in those systems
budgets. Rewards given by merit-pay systems should be
The two agencies used as samples in this study offer meaningful to each employee and group. Otherwise,
public administrators pay-for-performance according to those rewards do not play a sufficiently powerful role in
the performance record and multilateral ratings, as in motivating employees to work hard (Risher, 2004). It is
most other departments. Pay-for-performance in these important for organizations and leaders to know what
two agencies is paid at the level of the individual not at kinds or amounts of rewards are meaningful to employ-
the level of the department. ees. In some situations, non-cash rewards may be more
appropriate than financial ones.
As shown in Figure 1, the vertical and horizontal
A DIAGNOSIS MODEL FOR SUCCESSFUL dimensions each have two variables. Logically, there is a
PAY-FOR-PERFORMANCE high possibility that variables in each dimension have a
positive correlation. Leaders and organizations are likely
Public sector agencies have limited experience with to find or design meaningful rewards through employee
pay-for-performance and the effectiveness in the public participation. By contrast, without participation and
sector has been doubted (Perry, 1992). In 1993, the fed- decentralization they may have a hard time in meeting
eral government in the United States abandoned a merit- the expectations of employees. Also, they are more like-
pay system for middle managers originally introduced ly to apply standardized rules, whose inappropriateness
as part of the Civil Service Reform Act of 1978.3 In may become the cause of failure.
January 2007 Do Lim Choi 33

Figure 1. A Simplified Diagnosis Model for Pay-for-Performance

Source: Choi (2004).

Horizontal Dimension: pay-for-performance where leaders play vital roles in

Measurability and Appraisal forming a sound appraisal atmosphere. The appraisal
process should not be distorted by the influence of
Measurability means the extent of the possibility to seniority and informal groups. Therefore leaders play a
measure tasks or goals. Outputs and outcomes can be key role in creating an appropriate culture for pay-for-
considered in order to measure individual or organiza- performance.
tional performance. Outputs are measures of the quanti- In addition to measurement systems, Behn (2003)
ty of services provided or the quantity of service that emphasizes that goal setting for performance will have
meets a certain quality requirement. Outcomes are mea- an impact on results. To implement this motivational
sures of the results that occur because of services pro- strategy, the leadership of an agency needs to give peo-
vided (Melkers & Willoughby, 2005). Relatively speak- ple a significant goal to achieve and then use objective
ing, measuring outputs is easier than measuring out- performance measures, including interim targets. To
comes. The focus of many performance-based efforts motivate people to work harder, public sector managers
involves the evolution from basic activity-level mea- need measures of outputs to compare with production
sures to more meaningful outcome and impact measures targets, rather than measures of outcome.
(Melkers & Willoughby, 2005). As in the vertical dimension, two variables located at
Measurability will be one of the most crucial factors the same horizontal dimension, appraisal and measura-
influencing fair appraisal in pay-for-performance sys- bility, are also likely to have a high possibility of a posi-
tems. Bruijn (2001) mentioned the four most frequently tive correlation. Without adequate measurement sys-
discussed functions of performance measurement. tems, fair and valid appraisal is unlikely to be achieved.
Appraisal is one of them. To offer merit-pay according Equipment and resources for sound measurement sys-
to the extent of performance, there should be adequate tems are the foundation of fair assessment.
performance measurement systems. Organizations too
frequently initiate contingency pay plans without tested Diagonal Dimension: Participation &
and validated measurement systems (Perry, 1995). Espe- Measurability vs. Reward & Appraisal
cially, public agencies that lack experience with perfor-
mance measurement need to invest significant time and The diagonal dimension shows the prerequisites for a
effort in the design and administration of measurement pay-for-performance system. Employee participation
systems. will be required to make the reward mechanism mean-
As explained in the previous section, fair and valid ingful to employees. Also, a fair and valid appraisal will
appraisal of employee performance is not easy work for need adequate performance measurement systems.
supervisors. It is also a critical factor in the success of Therefore, participation and measurability will be the
34 Development of a Diagnosis Model for Pay-for-Performance in the Public Sector and Its Application Vol. 11, No. 2

foundation of meaningful rewards and fair appraisals. RESEARCH METHODS

Participation and measurability will be good tools for
offering meaningful rewards and fair appraisals. The fol- To measure all independent variables in this study, 5-
lowing cases illustrating different combinations of these point Likert-type scales ranging from strongly dis-
variables offer five different examples of pay-for-perfor- agree to strongly agree were employed. All questions
mance systems. for independent variables were developed based on the
concept and critical issues of each variable or reused
Type A (low participation, no meaningful rewards, from previous research. In addition, open-ended ques-
low measurability, and unfair appraisal): Public agen- tions were used to examine the detailed reasons of the
cies in a situation of this type are the worst case for pay- thoughts of subjects on some important key questions.
for-performance. Without devoting significant time and There were four dimensions used in this study: par-
effort, these organizational pay-for-performance systems ticipation, rewards and motivation, measurability, and
will fail. appraisal. Multiple questions were developed in order to
measure each dimension. The relevant hypothesis exam-
Type B (high participation, meaningful rewards, low ined in this study will be presented after the explanation
measurability, and unfair appraisal): Above all, public about each dimension and the questions asked.
agencies belonging to Type B should make efforts to For the participation dimension, three questions were
develop adequate measurement systems and then, based used. One was a question about participation in the
on these measurement systems, implement a fair process of general decision-making, another about the
appraisal of performance. same in the process of general decision-making on pay-
for-performance, and the other about participation in the
Type C (low participation, no meaningful rewards, process of decision-making on the type of pay-for-per-
high measurability, and fair appraisal): Public agencies formance.
belonging to Type C have a relatively good environment
in which to introduce merit-pay systems. Making Hypothesis 1: In public sector organizations,
employees feel positive about rewards through partic- employees who enjoy a high degree of participation in
ipation will result in the success of the pay-for-perfor- the decision making process will be satisfied with merit-
mance system. Assigning sufficient resources and sur- pay systems.
veying the needs of employees are both necessary to
achieve a better situation. Relevant to rewards and motivation, four questions
were introduced. One was a question about the satisfac-
Type D (middle level in all variables): Many pubic tion with rewards offered from pay-for-performance sys-
organizations will probably be located not in extreme tems. Another was a question about the attractiveness of
cases like Type A, B, and C, but somewhere around rewards. The third one asked was whether one making
Type D. Public agencies belonging to this type should more effort is also related to the attractiveness of
make efforts to improve all variables but, especially as rewards. If employees do not place more effort to get
noted in diagonal dimension, pay close attention to par- those rewards from pay-for-performance systems, they
ticipation, and measurability. These two factors will be will not be attractive. The final question asked is
the foundation for fair appraisals and meaningful whether employees who show excellence in work
rewards. receive correspondent rewards through pay-for-perfor-
mance. If employees do not get correspondent merit-pay
Type I (Ideal Type): Public agencies of this type are they will be de-motivated.
in the best possible situation for implementing pay-for-
performance. Organizations and leaders should not Hypothesis 2: Organizations where employees
assume that there is no more work required to make the enjoy a high degree of participation in the decision mak-
merit-pay systems better. What is still needed to keep in ing process will offer meaningful rewards to employees
mind is system maintenance with continuous employee producing high performance.
feedback.
For measurability, three questions were adminis-
January 2007 Do Lim Choi 35

tered. One question was asked respectively about the FINDINGS

quantitative characteristics of work, the comparability of
performance, and the short-term based measurability. Basic Statistics
Finally, four questions were posed in the dimension
of appraisal. One related to the satisfaction of perfor- This section reports basic information and statistics
mance appraisal. One concerned the fairness of the for further research such as response rate, demographic
appraisal. The rest of them were from the criteria of characteristics, and Cronbach alpha.
appraisal including patronage and the relationship
between pay and performance. Response Rate

Hypothesis 3: High measurability for job perfor- Two agencies were selected for this research and a
mance will provide organizations with good tools for survey was administered in 2006. The total number of
valid and fair appraisals. employees in those agencies was 259, and 240 surveys
were distributed to the employees working for those
Hypothesis 4: Although high measurability is neces- agencies. Of those distributed, 145 questionnaires were
sary for valid and fair appraisals it is not by itself a satis- collected with one incomplete and disregarded. The144
factory condition. usable surveys (that yielded a response rate of 60%)
were used for analysis in the research.
Hypothesis 5: Organizations using merit-pay sys-
tems that are not successful will be in a situation of low Descriptive Statistics
employee participation or low measurability, because
these are fundamental factors for the success of pay-for- Detailed descriptive statistics are reported in Table 2.
performance. As shown in Table 2, 57.04% of the respondents
were male and 42.96% female. Most of the employees
Two agencies were selected to sample and to apply were in the 30s (43.06%) or 40s (34.03%). Regarding
the diagnosis model in the practical arena. In terms of education, 63.64% were university graduates and
job characteristics, agency organizations fit the purpose 26.52% had high school diplomas. Of the respondents,
of this study and are operated within the patterns most 36.81% had worked fewer than 10 years in government,
similar to market-oriented organizations. Questionnaires 47.22% for 11 to 20 years, and 15.97% for more than 21
were distributed to all public servants of grade 5 and years. Finally, various grades and series were included
lower who receive pay-for-performance in the agencies. for the respondents.
To analyze the data, quantitative methods were The surveys were distributed to almost all employees
employed, such as variance, correlation, regression, and of the two agencies and collected at a rate of 60%. The
path analyses. First, frequency analysis was used to collected data represents the overall population of the
examine demographic characteristics and basic respons- two agencies.
es to each question. Second, the study employed Cron-
bach alpha and means tests to check the unidimension- Credibility and Correlation
ality of multiple questions and to respectively take a
look at the initial shape of the variables. Third, a corre- Table 3 shows means, standard deviation, Cronbach
lationship analysis was used to see the relationships alpha, and correlation coefficients. Through this table,
among variables. Fourth, to test the means difference one can obtain basic information about the thoughts of
between groups, variance analysis was administered. respondents for all variables and to a degree guess the
Finally, the study used regression and path analysis to results. As shown in the appendix, multiple questions
examine the causal effects between independent and were asked to measure one variable. A Cronbach alpha
dependent variables. was used to measure the unidimensionality of multiple
questions. Most of the variables show reliability coeffi-
cients of 0.7 or higher, and only the alpha of the partic-
ipation variable is 0.6 which can be acknowledged as
unidimensionality.
36 Development of a Diagnosis Model for Pay-for-Performance in the Public Sector and Its Application Vol. 11, No. 2

Regarding correlation, all variables are statistically negative feelings or perceptions about all variables rele-
correlated with each other and are all positively correlat- vant to performance-based pay. Measurability shows the
ed. Meaningful rewards and satisfaction with merit pay highest mean (2.85) that reflects on the work character-
show the highest coefficient (0.71). Fair appraisal and istics of the agency, but the result indicates the level of
satisfaction with merit-pay show the second strongest measurability is not enough. Satisfaction about perfor-
correlation coefficient (0.68). Measurability and fair mance-based pay shows the lowest mean value (2.60). It
appraisal have the lowest correlation coefficient (0.17). is concluded that public servants in the two agencies are
As can be seen in the means, the respondents showed unsatisfied with performance-based pay.

Table 2. Demographic Characteristics

Frequency (total no.: 144)*

Classification
number percentage
male 81 57.04
gender
female 61 42.96
20s 22 15.28
30s 62 43.06
age
40s 49 34.03
50s 11 7.64
high school 35 26.52
education four-year university 84 63.64
post-graduate 13 9.85
1 to 10 years 53 36.81
work years 11 to 20 years 68 47.22
over 21 years 23 15.97
grade 5 and higher 26 29.21
grade
grade 6 and lower 63 70.79
general 19 18.81
special** 31 30.69
series contract 17 16.83
functional 20 19.80
others 14 13.86
*: Each demographic characteristic has a frequency missing which indicates the number of respondents that did not mark that question.
**: It indicates public officials of grade 6 and lower who work in positions where special skills are required and belong to a non-career track.

Table 3. Means, Standard Deviation, Credibility, Correlation (n=144)

Correlation coefficient
variable mean std.
satisfaction measurability participation meaningful rewards
satisfaction 2.60 0.73 0.9
measurability .17* 2.85 0.68 0.7
participation .48** .21* 2.82 0.59 0.6
meaningful rewards .71** .18* .53** 2.79 0.57 0.7
fair appraisal .68** .17* .52** .59** 2.79 0.62 0.8
*: significant at 0.05 **: significant at 0.01
January 2007 Do Lim Choi 37

Hypothesis Test work hard as compared with coworkers who do not

work as hard because of the lack of resources and unfair
Vertical Dimension: Participation and Meaningful appraisal. Second, they do not agree with the idea of get-
Rewards ting different levels of salary when they work together
on the same task.
Hypothesis 1: In public sector organizations,
employees who enjoy a high degree of participation in Horizontal Dimension: Measurability and Appraisal
the decision-making process will be satisfied with merit-
pay systems. Hypothesis 3: High measurability for job perfor-
As shown in table 4, there were significant differ- mance will provide organizations with good tools for
ences in the level of satisfaction between public servants valid and fair appraisals.
with high participation and low participation. Compared The finding in Table 5 supports hypothesis 3 when
to low participation public servants, those with high par- there were statistically significant differences in the per-
ticipation were significantly higher in the satisfaction ception of appraisal fairness between agency employees
and this strongly confirms hypothesis 1. who believe that work is highly measurable and those
who do not. Compared to agency employees who do not
Hypothesis 2: Organizations where employees believe that work is highly measurable, those who do
enjoy a high degree of participation in the decision-mak- thought that the performance appraisal was fair and this
ing process will offer meaningful rewards to employees strongly confirms hypothesis 3.
producing high performance. Mean value, remained negative (2.93). That means
Table 4 shows that there were significant differences agency employees were still not satisfied with the results
in reward meaningfulness between public servants with of the performance appraisal, although the work was
high participation and low participation. Compared to measurable. As a result agencies will need to use some-
public servants who show low participation in the orga- thing more than just the measurability of performance to
nizations, those with high participation were more satis- achieve a fair appraisal.
fied with the performance-based rewards and this Some reasons that a fair appraisal is difficult to be
strongly confirms hypothesis 2. achieved in the public sector were exposed through the
Respondents who negatively answered the question open-ended questions. First, respondents thought that
on the meaningfulness of pay-for-performance were the subjective or personal opinion of the appraisers is a
asked why they were not satisfied with rewards through factor in the process of appraisal. Second, they insisted
performance evaluation. The main reasons were as fol- that it is not easy to measure the work, or that it is diffi-
lows. First, they mentioned the equity issue, where pub- cult to quantify work in government.
lic servants do not receive rewards to the extent that they

Table 4. Satisfaction Level and Reward Meaningfulness According to Participation

High Participation (n=36) Low Participation (n=79)

Dependent Variable t p
Mean Std. Mean Std.
satisfaction 3.18 .61 2.31 .67 -6.57 .000
meaningful rewards 3.24 .48 2.54 .50 -7.05 .000

Table 5. Fairness, seniority, and patronage in performance appraisal according to measurability

High Measurability (n=46) Low Measurability (n=66)

Dependent Variable t p
Mean Std. Mean Std.
fair appraisal 2.93 .71 2.67 .59 -2.18 .03
seniority 2.78 .87 2.77 .86 -.06 .95
patronage 3.15 .99 2.94 .99 -1.12 .26
38 Development of a Diagnosis Model for Pay-for-Performance in the Public Sector and Its Application Vol. 11, No. 2

Hypothesis 4: Although high measurability is neces- Hypothesis 5: Organizations using merit-pay sys-
sary for valid and fair appraisals, it is not a singularly tems that are not successful will be in the situation of
satisfactory condition. low employee participation or low measurability,
In hypothesis 3, it is indirectly known that a high because these are fundamental factors for the success of
measurability is not a singularly satisfactory condition pay-for-performance.
for fair appraisal. To find better evidence, the reflection To confirm hypothesis 5 a regression analysis
of seniority and patronage in the process of performance between satisfaction in merit pay and four potential
appraisal was examined. influence factors was implemented. The result is shown
As shown in Table 5 there were no significant differ- in Table 6. Appraisal fairness and reward meaningful-
ences in the reflection of seniority between agency ness emerged as statistically significant. The findings
employees who believe that work is highly measurable explain that those two variables are direct and influential
and those who do not. Regardless of performance mea- factors that affect the satisfaction level in performance-
surability the means of two groups in seniority reflection based pay. Participation and measurability were not sta-
were the same. Both groups thought that seniority was tistically significant in this regression model.
reflected in the process of performance appraisal. Figure 2 shows that participation and measurability
The reflection of patronage in the process of perfor- are fundamental variables for a successful performance-
mance appraisal was also studied. The finding in Table 5 based pay system. Participation and measurability influ-
shows that there were no significant differences in the ence respectively reward meaningfulness and appraisal
reflection of patronage between agency employees who fairness and this confirms hypothesis 5.
believe that work is highly measurable and those who
do not. Without regard to measurability, the means of
agency employees in both groups showed almost the DISCUSSION
same value around neither agree nor disagree.
The findings showed that high measurability is not as a In this section, practical policy ideas will be suggest-
singularly satisfactory condition for a valid and fair ed based on the results of the analyses.
appraisal, albeit a necessary one and this confirms hypoth- Employee participation will be an important factor
esis 4. to increase the level of satisfaction regarding perfor-
mance-based pay in that an organization ought to make
Diagonal Dimension: Participation & Measurability more efforts to establish participative mechanisms and
vs. Reward & Appraisal. a participative culture. Without participation and con-
sensus it will be hard to make a success of perfor-

Table 6. Relationships between Satisfaction in Performance-based Pay and Influential Factors

Dependent Variable: Satisfaction in merit pay

Independent variable
Beta Standard error t
participation .02218 .08290 .33
measurability .01865 .05878 .34
appraisal fairness .38904 .08153 5.57**
reward meaningfulness .46866 .08967 6.64**
F Value: 52.86 R-Square: 0.6121 Adj R-Square: 0.6005 **: Significant at 0.01

Figure 2. Path Analysis

January 2007 Do Lim Choi 39

mance-based pay. Employees should be encouraged to The fifth issue is work measurability in the public
take part in the process of decision-making about merit- sector. Work measurability is essential for the accep-
pay itself, types of rewards, measurement of perfor- tance by employees in the objectivity toward the results
mance, and appraisal. of performance, but most of the public servants realized
Second, changes in the attitude and culture public that it is difficult to measure the work or performance.
servants will be required to accomplish a fair appraisal Conditions under which performance measurement is
of performance. Agencies should focus not only on problematic are shown in the following observations
organizational structures or systems, but also on the (Bruijn, 2002). First, an organization has obligations
change in organizational culture. Existence of rational and is highly value-oriented. Second, products are mul-
rules and procedures to measure employee performance tiple. Third, an organization is process-oriented. Fourth,
does not secure fair performance appraisal as long as products are generated together with others. Fifth, prod-
irrational customs like patronage exist. Respondents ucts are interwoven. Sixth, quality is not definable in
insisted that the subjective perspective of appraisers is performance indicators. Seventh, products are various.
involved in performance appraisal. Finally, the environment is dynamic. In this context, the
Third, it is incumbent on agencies to make an effort excessive emphasis on work measurability can cause
to increase the level of satisfaction regarding pay-for- the replacement of goals and means. Bruijn (2002)
performance. A negative response was identified insists that performance measurement veils actual per-
through structured survey questions on pay-for-perfor- formance.
mance. Through open-ended questions, some reasons The sixth issue is whether agencies are ready to
for this negativity were exposed. Respondents indicated employ performance-based pay systems and whether
as reasons, unfair appraisal, lack of resources, difficulty governments are ready to introduce performance-based
of measurement, encouragement of competition, and pay systems. According to the results of this study, it
inappropriateness of appraisal unit. There seem to be will be hard to answer yes. The public sector needs to
many obstacles to making pay-for-performance success- invest more time and effort in order to make merit pay
ful in agencies. systems that are satisfactory to public servants. Since the
Fourth, agencies should consider the attractiveness of introduction of the pay-for-performance system, almost
merit-pay. As indicated above, respondents are not very a decade has passed, but the system still has not stabi-
attracted to merit-pay in the organization. The reasons lized (refer to Figure 3). Above all, for stabilizing the
are as follows: they feel the amount of merit-pay is not system, agencies have to pay attention to the fundamen-
enough to be satisfactory, the merit-pay system encour- tal variables, participation, and measurability, for suc-
ages competition among coworkers, and that compensa- cessful performance-based pay systems. Then they
tion is mainly based on financial rewards. should move to the other two to achieve a fair appraisal

Figure 3. Current Status of Agencies on the Simplified Diagnosis Model

40 Development of a Diagnosis Model for Pay-for-Performance in the Public Sector and Its Application Vol. 11, No. 2

and meaningful rewards. REFERENCES

Lastly, agencies must comprehensively consider the
need for performance-based pay in the public sector in Ammons, D. N. and Rodriquez, A. 1986. Performance
whether agencies or governments need merit pay to Appraisal Practice for Upper Management in City
increase productivity and to examine if there are other Governments. Public Administration Review,
methods. 46(5): 460-467.
Behn, Robert D. 2003. Why Measure Performance?:
Different Purposes Require Different Measures.
CONCLUSION
Public Administration Review, 63(5): 586-606.
In this section the brief results and implications of Bruijn, Hans De. 2002. Managing Performance in the
this study will be summarized so as to lead to future Public Sector. New York: Routledge.
research issues that need to be discussed. Choi, D. L. 2004. Merit-Pay System and Motivation: A
The brief results of this study are as follows. First, Diagnosis Model. The Research Review of Regional
employees who enjoy a high degree of participation in Development, 16: 169-183 (in Korean).
the decision-making process are satisfied with merit-pay
systems. Second, organizations where employees enjoy Crewson, Philip E. 1997. Public-Service Motivation:
a high degree of participation in the decision-making Building Empirical Evidence of Incidence and
process offer meaningful rewards to employees produc- Effect. Journal of Public Administration Research
ing high performance. Third, high measurability for job and Theory, 7(4): 499-518.
performance provides organizations with strong tools Dakahashi, Shunskae 2000. Beyond Pay for Perfor-
for valid and fair appraisals. Fourth, although high mea- mance. Translated by T. J. Jung & H. K. Ham in
surability is necessary for valid and fair appraisals, it is Korean. Muhan Co.
not by itself a satisfactory condition. Finally, participa-
Deci, E. L. 1975. Intrinsic Motivation. New York:
tion and measurability are fundamental factors for the
Plenum.
success of pay-for-performance.
The main limitation of this study is the omission of Flannery, T. P., Hofrichter, D. A., & Platten, P. E. 1996.
some more influential factors that are relevant to pay- Dynamic Compensation for Changing Organiza-
for-performance in the simplified diagnosis model. In tion: People, Performance, & Pay. The Free Press.
addition many recent studies have been implemented Ha, M. S. 2004. Problem Diagnosis and Alternatives for
under the assumption that performance-based pay was Performance Management Systems in Government:
given. More studies about the justification of merit pay In Comprehensive Perspective. The Korea Institute
and the preparation of will be required in the future. of Public Administration Journal, 13(1): 235-273.
Hatry, Harry P. 1999. Performance Measurement Princi-
NOTES ples and Techniques: An Overview for Local Gov-
ernment. In Public Sector Performance (Ch.22),
* This study was financially supported by research fund of edited by Richard C. Kearney & Evan M. Berman.
Chungnam National University in 2005. Westview Press.
1. In this study, pay-for-performance and merit pay have the
Hatry, H. P. 1999. Performance Measurement: Getting
same meaning and are used interchangeably.
Results. Washington, DC: Urban Institute.
2. Bruijn (2002) suggests that there are four most frequently
mentioned functions of performance measurement: trans- Herbert, A. 2004. Transforming the public service in a
parency, learning, appraising, and sanctions. world without boundaries. Donald Stone Lecture at
3. The change in policy gained new momentum when the American Society for Public Administration Nation-
Department of Defense and Homeland Security were al Conference, ASPANET This Week, May 10.
authorized to develop specific pay systems. Pay for per-
formance is also the central change in the way the mem- Hwang, C. Y. 2003. A Study on the Relationship
bers of the Senior Executive Service (SES) are being between Organizational Culture and Performance
compensated (Risher, 2004). Management. Korea Policy studies Review, 12(2):
January 2007 Do Lim Choi 41

283-320. OECD.
Hwang, S. W. 2003. A Thought on Performance Payment Oh, S. H. 2003. A Study on the Problems and Improve-
System in the Public Sector. The Korea Institute of ment of Local Government Performance-Based Pay
Public Administration Journal, 12(4): 123-147. System. The Korea Institute of Public Administra-
Ingraham, P. W. 2005. Performance: Promises to Keep tion Journal, 12(4): 219-247.
and Miles to Go. Public Administration Review, Pearce, J. L., W. B. Stevenson, and J. L. Perry. 1985.
July/August, 65(4): 390-395. Managerial Compensation Based on Organizational
Kang, H. S. and Kwon, Y. S. 2004. Public Employees Performance: A Time-series Analysis of the Impact
Acceptability on the Results of Organizational Per- of Merit Pay. Academy of Management Journal,
formance Evaluation: Focused on Systematic Fac- 28 (June): 261-278.
tors of Performance Evaluation. Korean Public Perry, J. L. 1992. The Merit Pay Reforms. In The
Administration Quarterly, 16(4): 789-811. Promise and Paradox of Civil Service Reform,
Kilpatrick, F. P., Cummings, M. C., and Jennings, M. K. edited by P. W. Ingraham & D. H. Rosenbloom.
1964. The Image of the Federal Service. Washing- Pittsburgh University of Pittsburgh Press.
ton, D. C.: Brookings Institution. _______. 1995. Compensation, Merit Pay, and Motiva-
Kim, J. H. 1999. Commentaries: The Executive Agency tion. In Public Personnel Administration: Prob-
in Korean. Journal of Governmental Studies, lems and Prospects (Ch.8), edited by Stenven W.
11(1): 177-194. Hays & Richard C. Kearney. Prentice-Hall.

Lee, H. T. 2005. Reaction and Political Issues on Pay- Rainey, Hal G. 1997. Understanding and Managing
for-Performance in Local Governments. Korean Public Organizations. San Francisco: Jossey-Bass.
Public Administration Quarterly, 17(1): 165-186. Risher, Howard. 2004. Pay for Performance: A Guide
Lee, H. U. 2005. Reexamining Competition And Coop- for Federal Managers, IBMs Center for The Busi-
eration as Factors of Organizational Objectives ness of Government.
Achievement. Korean Public Administration Roberts, A. 1997. Performance-Based Organizations:
Review, 39(2): 87-110. Assessing the Gore Plan. Public Administration
Meyer, H. H. 1975. The Pay for Performance Dilemma. Review, 57(6) : 465-478.
Organization Dynamics, 3(Winter): 39-50. You, H. 1999. A Study on Executive Agencies in Korea.
Melkers, Julia and Willoughby, Katherine. 2005. Mod- Korean Journal of Public Administration, 37(1):
els of Performance-Measurement Use in Local 57-72.
Governments: Understanding Budgeting, Commu- Young, I. Phillip. 2003. The Trouble with Pay for Perfor-
nication, and Lasting Effects. Public Administra- mance. American School Board Journal, Novem-
tion Review, 65(2): 180-190. ber: 40-42.
Menzel, Donald C. 1993. The Ethics Factor in Local
Government: An Empirical Analysis. In Ethics and
Public Administration(Ch.10), edited by H. George Do Lim Choi is an assistant professor at the Department
Frederickson. New York: M. E. Sharpe. of Public Administration at Chungnam National Univer-
sity. Teaching and research interests include public
Nunberg, Barbara. 2002. Exporting Administrative ethics, human resources management, and organization-
Excellence: Adapting Advanced-Country Models al theory. His research has appeared in various journals
to Developing Contexts. In Public Personnel Man- including Journal of Public Administration Research
agement: Current Concerns, Future Challenges and Theory and International Review of Public Admin-
(Ch.5), edited by Carolyn Ban & Norma M. istration. E-mail: dmchoi@cnu.ac.kr
Riccucci. Longman.
OECD. 1993. Private Pay for Public Work: Perfor- Received: August 29, 2006
mance-Related Pay for Public Sector Managers. Accepted after one revision: December 29, 2006