Meyer Sahling Schuster Mikkelsen - What Works in Civil Service Management
Meyer Sahling Schuster Mikkelsen - What Works in Civil Service Management
Meyer Sahling Schuster Mikkelsen - What Works in Civil Service Management
Report for the UK Department for International Development (DFID), Feb. 2018
Contact details of authors:
[email protected] [email protected] [email protected]
Acknowledgments
Acknowledgments
This report draws on results from an international survey of 23,000 civil servants in ten
countries in Africa, Asia, Eastern Europe and Latin America. The authors gratefully
acknowledge the financial support from the British Academy – UK Department for
International Development (DFID) Anti-Corruption Evidence Programme
(http://www.britac.ac.uk/node/4662/).
Implementing a survey of civil servants across ten countries and four continents was
only possible thanks to the collaboration of a great many colleagues and government
counterparts. Our thanks extend to, first, our academic collaborators, who took the
lead in or collaborated in implementing the survey in their respective country of
expertise: Adam Harris (Uganda), Brigitte Seim (Malawi), Rachel Sigman (Ghana),
Tiina Randma-Liiv and Cerlin Pesto (Estonia), Izabela Correa (Brazil), Shreekrishna
Shrestha (Nepal), Hamit Qeriqi (Kosovo), Ansi Shundi (Albania) and Taiabur Rahman,
Kazi Marful Islam and Ahmed Shafiql Huque (Bangladesh). Our academic country
collaborators have also lead-authored a series of country reports on civil service
management, which complement this cross-country report with country-specific
findings and are (or will shortly become) available for download from the project’s
website: http://www.britac.ac.uk/node/4662/. These reports also duly acknowledge the
dozens of research assistants whose excellent research assistance was central to
running and analysing the survey.
Further, we are grateful for the support and authorization of the survey by government
institutions in the ten countries studied. Civil service surveys not only provide an
evidence basis for civil service reform. They also put a transparency spotlight on
public employment. We appreciate the governments’ support for this endeavour.
Finally, we would like to thank the over 23,000 civil servants who took the time to
complete the survey and share their experiences in public service. Without them,
there would have been no report.
The views and opinions expressed in this report are those of the authors and do not
necessarily reflect the views and opinions of any government, funding agency or
university.
Executive Summary
“Civil service
reforms require
Executive Summary
tailoring to the Civil servants are central to effective governance in developing countries. They deliver
realities of each essential services to citizens, commission infrastructure, regulate economic activity
and engage in diplomacy with foreign countries – to name just a few tasks. This puts a
institution”
premium on understanding how to manage civil servants in developing countries
effectively. Yet, to-date, there are scarcely any quantitative studies which deliver
robust findings across developing countries – let alone regions – on what works in civil
service management. To address this gap, this report draws on data from an original
survey of 23,000 civil servants in ten countries in Africa, Asia, Eastern Europe and
Latin America – the, to our knowledge, largest original cross-country survey of civil
servants ever conducted in the developing world.
Drawing on this data, the report assesses the effects of a range of civil service
management practices – from recruitment to promotion, pay and performance
management practices – on the attitudes and behaviour of civil servants. To
understand the desirability of these practices, our survey covers a spectrum of civil
servant attitudes and behaviour which are core to civil service effectiveness: work
motivation, job satisfaction, public service motivation, commitment to remaining in the
public sector, performance and integrity. With these indicators, we can identify which
civil service management practices tend to have positive effects and which do not –
thus providing a foundation for evidence-based civil service reform designs.
What can be learned from the data? First of all, that effects of civil service
management practices need to be understood within countries and institutions. The
attitudes and behaviour of civil servants sharply vary across and within countries and
institutions. The resulting pattern bears little resemblance to conventional wisdoms
about developing country states. Many prior studies had construed them as
dichotomies between ‘islands of excellence’ and seas of mediocrity. Top performers
and basket cases certainly exist in our data. Most institutions, however, are neither.
Instead, they sit in between. Gradual differences rather than dichotomies between
poor and strong performance mark most institutions in developing country civil
services.
Moreover, institutions (and civil servants) which score highly in one attitude or
behaviour (e.g. work motivation) often do not do so in another (e.g. commitment of
civil servants to remain in public sector). Institutions may thus have strengths in some
dimensions of civil servant behaviour and attitudes, while having weaknesses in
others.
These findings underscore that civil service reforms ultimately require tailoring to the
realities of each institution – and, at times, the realities of each unit or group of
professionals within institutions. To tailor to local realities thereby requires an
Executive Summary
1 Many civil service reforms beyond these four may, of course, be beneficial in any given
country. We selected these four given their relatively consistent effects across countries.
Executive Summary
Contents
Acknowledgments _____________________________________________________ i
Executive Summary ___________________________________________________ ii
I. Introduction ________________________________________________________ 1
II. The Nature of civil servants in developing countries _________________________ 6
Insight #1: The attitudes of civil servants are dynamic ___________________ 8
Insight #4: Good attitudes need not coincide _________________________ 10
Insight #3: Attitudes and behaviour vary across and within countries and
institutions ___________________________________________________ 12
Insight #4: Most institutions are neither 'islands of excellence' nor basket cases
____________________________________________________________ 14
III. Civil service management in developing countries: what works? _____________ 17
Lesson #1: Depoliticize civil service management _____________________ 18
Lesson #2: Curb nepotism in civil service management _________________ 27
Lesson #3: Ensure that performance matters in civil service management __ 36
Lesson #4: Pay enough to retain (more) motivated civil servants _________ 49
IV. Implications for civil service reform ____________________________________ 59
Appendix __________________________________________________________ 61
Bibliography ________________________________________________________ 77
Pg. 01 Introduction
“This report
draws on the
I. Introduction
largest original
international Making civil services work: an international research project
survey of civil
Civil servants are central to effective governance in developing countries. They deliver
servants to-date,
essential services to citizens, commission large-scale infrastructure, regulate
with over 23,000
economic activity and engage in diplomacy with foreign countries – to name just a few
respondents in
roles. Yet, in many developing countries, civil servants do not consistently take on
ten countries in
these roles effectively. Instead, civil services are marked by service delivery failures
four developing
and corruption. In response, donor organizations and governments in developing
regions.”
countries have recurrently sought to reform civil services. The World Bank (2008), for
instance, lent US$422m per year for this purpose between 2000 and 2006. The track
record of these reforms has been far from stellar, however. The World Bank’s lending,
for instance, had no measurable impact on civil services (World Bank, 2008). Other
aid organizations raise similar concerns (e.g. DFID, 2011).
One important hindrance to reform has been the absence of rigorous evidence on
how to manage civil servants effectively in developing country contexts. In fact, some
observers go as far as noting that “we do not really know what we are doing” in civil
service reform (Brösamle, 2012). While this might be an exaggeration in light of some
recent studies (see e.g. Rogger, 2017), it is clear that, to-date, there are scarcely any
quantitative studies which deliver robust findings across developing countries – let
alone regions – on what works in civil service management in developing countries.
This report seeks to help address this gap. It draws on results from the – to our
knowledge – largest original cross-country survey of civil servants conducted to-date,
with responses from 23,000 civil servants in ten countries in four developing regions:
Ghana, Malawi, Uganda, Brazil, Chile, Estonia, Albania, Kosovo, Bangladesh and
Nepal. The underlying research project was led by Jan-Hinrik Meyer-Sahling
(University of Nottingham) and Christian Schuster (University College London) and
Pg. 02 Introduction
Our cross-country survey gathers data on civil servants’ experiences and perceptions
of civil service management practices on the one hand, and civil servants’ attitudes
and behaviour on the other. The civil service management practices covered in the
survey include recruitment, pay, promotion, performance evaluation and career
management of civil servants. Concurrently, the survey captures a broad spectrum of
attitudes and behaviour of civil servants – their work motivation, job satisfaction, public
service motivation, commitment to remain in the public sector, performance and
integrity.
We hope this analysis provides an evidence basis and starting point for governments
and donors hoping to design improvements in civil service management.2
2 This report is based on cross-sectional analyses of perception-based survey data, which can
provide important insights, but is not without limitations. As such, we hope that its findings are
treated as a basis to discuss potential improvements to civil service management practices –
rather than a be-all and end-all guide to civil service reform.
3 Legally, the scope of civil services varies across countries. To ensure comparable samples,
we did not follow legal definitions but rather the aforementioned common usage of the ‘civil
service’ in our survey sampling.
Pg. 03 Introduction
top to the bottom of the administrative hierarchy. Most surveyed civil servants were
“Our statistical employed in the capital city of the respective countries.
analyses can
provide The survey was conducted with authorization and support from central governments in
guidance on the ten countries. How institutions and civil servants were sampled within the civil
service in each country varied somewhat, due to differences in the survey mode,
which
scope of government support and availability of survey population data.
management
practices tend to In five countries, respondents completed the survey online, based on governments
have positive (or holding email records of civil servants. In three of these countries, the survey was sent
negative) effects to all ‘civil servants’ (Kosovo, Estonia and Albania), except officials employed in
on the attitudes defense ministries and their subordinated organizations. In two of these countries, the
and behaviour of survey was sent to all civil servants in eleven (Chile) and fourteen (Brazil) central
civil servants.” government institutions. Response rates varied across countries (see Appendix A.1).
To enhance representativeness and subject to data availability, a subset of our
country reports includes survey weights (see, e.g., Schuster, Meyer-Sahling,
Mikkelsen, Gonzalez Parrao, 2017).
In the five remaining countries, the survey was conducted in-person. Civil servants
were selected through informal quota sampling and chain referral, with a view to
obtaining responses from civil servants in a variety of positions, institutions and
functions. The survey thereby extended to 48 (Ghana), 31 (Uganda), 62 (Malawi), 31
(Nepal) and 38 (Bangladesh) state institutions. Appendix A.1 contains further detail on
survey sampling and response rates.
The cross-country survey was translated (and, at times, back-translated) into local
languages where necessary. To ensure a comparable understanding of the wording of
our questions across our diverse range of countries and languages, the survey was
pre-tested in each country through a series of cognitive interviews with public
servants. The survey was iteratively revised in each country until cognitive interviews
with public servants suggested measures were understood as intended
.
Pg. 04 Introduction
The survey sample implies that our survey data can provide detailed insights into civil
servants in a wide range of positions and institutions in four different developing
regions. To illustrate, table 1 contains basic demographic data on our respondents.
Our respondents are virtually balanced in terms of gender; have, on average, over 13
years of experience in the public sector; are roughly 43 years old; are distributed
across levels of the administrative hierarchy, with a majority in technical-professional
positions; and tend to be university educated and on permanent contracts (with one in
five being on temporary contracts and not having a university degree, however).
While our survey thus covers a range of civil servants, institutions and countries, our
findings are not necessarily representative for each of the countries studied. Survey
mode and respondent sampling varied in each country, and we could not obtain data
to construct survey weights across countries. As a result, cross-country comparisons
in particular should thus be interpreted with care. They are only suggestive.
Chapter 3 turns to the core purpose of this report: what works in civil service
management? To shed light on this question, regression evidence is presented on the
effects of civil service management practices on the motivation, performance,
satisfaction, commitment and integrity of civil servants. The chapter derives four core
lessons about how to make civil services in developing countries work based on this
analysis. These lessons are structured around civil service management practices
which have positive effects in all or most of the ten countries surveyed.
The report concludes in chapter 4 with implications and policy recommendations for a
more motivated, committed and ethical civil service in developing countries.
Pg. 06 The nature of public servants in developing countries
#1: The attitudes of civil servants are dynamic: they progress over time
#2: Good attitudes need not coincide: civil servants and institutions which score
high in one attitude or behaviour (e.g. motivation to serve the public) need not
score high in another (e.g. motivation to work hard).
#3: Attitudes and behaviour vary sharply across and within countries and
institutions; at times, groups within institutions will differ more sharply in their
attitudes than the average civil servant across countries.
#4: Most institutions are neither ‘islands of excellence’ nor basket cases. Contrary
to conventional wisdom, they are instead situated on a gradual slope between
excellence and mediocrity.
For civil service reformers, the first finding is good news: if attitudes of civil
servants are dynamic, they can be shaped by civil service reforms. Findings 2 to
4 caution attention, however, to trade-offs and local contexts. As good attitudes
need not coincide, reforms can foster some at the expense of others; and, as
attitudes vary sharply between and within institutions, the effects of management
practices might plausibly do so as well. Reforms thus need to be tailored to the
realities of each institution. Understanding the attitudes and behaviour of civil
servants (which reforms, ultimately, seek to shape) should thus be the first step in
any civil service reform.
These are, of course, by no means the only desirable attitudes and behaviour of civil
servants. However, they are all attitudes and behaviour, which, according to prior
studies, matter for civil service effectiveness. Job satisfaction, a commitment to remain
in the public sector, work motivation, public service motivation and individual job
performance have all been associated with greater organizational performance (see
e.g. Cantarelli, Belardinelli & Belle, 2016; Kim, 2004; Ritz, 2009); as have integrity –
including (lower) corruption, clientelism and nepotism (see e.g. Gould and Amaro-
Reyes, 1983).
They also reflect a wide spectrum of civil service attitudes and behaviour. Thanks to
cross-country data on this wide spectrum, this report can identify several insights
about the nature of civil servants in developing countries, which studies focused on
single dependent variables or countries could often not.
Pg. 08 The nature of public servants in developing countries
Our data suggests, first, that how motivated, committed, satisfied and ethical civil
servants are varies significantly over time in public service. To illustrate, figures 2 to 3
plot the average job satisfaction and preference for remaining in the public sector
(public sector commitment) of civil servants in our sample by years of experience.
The figures suggest that these attitudes are dynamic. Average job satisfaction and
public sector commitment fall significantly in the first five years in public service, and
only recover to initial levels after 15 to 20 years of service. They reach their peaks
shortly before retirement. Public sector turnover – with the most dissatisfied staff
leaving – might well explain the recovery after five years; it does not provide a
convincing explanation for the initial fall, however. These shifts are also not merely
Pg. 09 The nature of public servants in developing countries
Figure 3. Proportion of civil servants preferring to work in the public sector by years of
public service experience
Pg. 10 The nature of public servants in developing countries
We observe a similar pattern at the institutional level. Some institutions score high in
some attitudes, but not others. Two rankings of attitudes and behaviour in the
surveyed state institutions in Chile are illustrative. Figure 4 compares the proportion of
civil servants who would prefer continuing to work in the public sector (over a private
sector job). Figure 5 compares the self-assessed performance of civil servants. There
is significant movement of institutions between the rankings. The civil service agency
(DNSC) scores among the top 3 institutions in terms of the preference of staff to
continue to working in the public sector, yet among the bottom 3 in (self-assessed)
performance, for instance. Institutions may thus have strengths in some dimensions of
behaviour and attitudes of civil servants, yet not others.
Pg. 11 The nature of public servants in developing countries
Figure 4. Proportion of civil servants preferring to work in the public sector by state
institution in Chile
Good attitudes and behaviour thus need not coincide, both for individual civil servants
and for state institutions. If they do not go together, however, then the desirability of
civil service management practices depends on how they affect the range relevant
attitudes and behaviour of civil servants. A core civil service management challenge is
thus the design of practices which shape the range of attitudes and behaviour of civil
servants positively – rather than a few at the expense of others.
Figure 6. Proportion of civil servants per country indicating that they are willing to
engage in corruption
Pg. 13 The nature of public servants in developing countries
Importantly, this variation extends not solely to countries. It is also observable across
institutions within countries – and within individual institutions. To illustrate, Chile is the
country with the highest job satisfaction on average in our sample. Yet, as illustrated in
figure 7, six other countries have a higher average job satisfaction than the institution
with the least job satisfaction in Chile. Within-country variation can thus trump cross-
country differences in the attitudes and behaviour of civil servants.
“Gradual
Insight #4: Most institutions are neither ‘islands of excellence’
differences rather nor basket cases
than dichotomies When comparing the average attitudes and behaviour of civil servants across
between poor and institutions, the resulting patterns bears little resemblance to conventional wisdoms
strong performers about developing country states. In qualitative and policy publications, these are often
construed as dichotomies between islands of excellence and seas of mediocrity (see
mark institutions
e.g. Leonard, 2008; Roll, 2014). Top performers and basket cases certainly exist in
in developing
our data. Most institutions, however, are neither. Instead, they sit in between – on a
country civil long and gradual slope of increasing or decreasing performance. Gradual differences
services.” rather than dichotomies between poor and strong performance thus mark developing
country civil services. Figures 8 and 9 illustrate this pattern with the examples of the
average job satisfaction and public service motivation for the (several hundred)
institutions in our sample. We observe it equally for our other indicators. What this
implies for civil service reforms is discussed next.
Figure 8. Job Satisfaction by institution (all countries; global mean set to zero)
Pg. 15 The nature of public servants in developing countries
Figure 9. Public service motivation by institution (all countries; global mean set to zero)
In assessing the desirability of civil service management practices, however, the data
point to an important note of caution. This desirability needs to be assessed against a
large range of attitudes and behaviour of civil servants. Those need not coincide, and
civil service management practices might thus well come with trade-offs, favourably
affecting one attitude or behaviour, while adversely affecting another. The next
chapter thus assesses the effects of civil service management practices on a range of
Pg. 16 The nature of public servants in developing countries
attitudes and behaviour of civil servants. As will be shown, civil service management
“Civil service practices often have heterogenous effects.
reforms can
require tailoring Finally, the effects of civil service management practices need to be understood within
to not only the countries and institutions. As noted, there is sharp variation in attitudes and behaviour
across and within countries and institutions – in a pattern which puts a damper on the
realities of each utility of the ‘islands of excellence’ metaphor. In light of the diversity of these attitudes
state institution, and behaviour across and within institutions, the effects of (many) civil service
but also within management practices are likely to vary. Cookie cutter civil service reforms are thus
state institutions unlikely to be effective. Rather, civil service reforms require tailoring to the realities of
each institution – and, at times, to the realities of each unit or group of professionals
the realities of
within (or across) institutions. Understanding these realities requires an appreciation of
each department not only existing management practices, but also the varied attitudes and behaviour of
or group of civil civil servants that are being shaped by them (for instance via staff surveys).
servants”
The next chapter takes this lesson on board and assesses both cross-country and
country-specific effects of civil service management practices. Moreover, several of
the country reports complementing this cross-country report also estimate effects of
civil service management practices in individual institutions (see, e.g. Schuster,
Meyer-Sahling, Mikkelsen, Gonzalez Parrao, 2017). The core lessons presented in the
next chapter are robust in most of these institutional-level analyses.
Pg. 17 What works in civil service management
The data in this chapter points to four core lessons for making civil services work
in developing countries:
These are, of course, not the only civil service management practices affecting
the motivation, satisfaction, performance and integrity of civil servants, nor do
these outcomes solely result from civil service management practices.
Nonetheless, our data suggests that they are worthwhile targets for civil service
reform attempts: we find support for these lessons and their effectiveness in
almost all of the ten countries studied.
Pg. 18 What works in civil service management
Figure 10 reports the proportion of civil servants who indicate that political connections
were at least somewhat (i.e. not not at all) important for their public sector jobs and
careers, that is, for their recruitment, pay and promotions.4 It shows that politicization
is not limited to the top of administrative hierarchies – where legitimate democratic
concerns with the political control of state institutions might warrant politicization (see
Kopecky et al., 2016). While politicization is most prevalent at the managerial level
(25% of civil servants for recruitment, 25% for promotions and 19% for pay), it also
occurs at the administrative support level (22%, 21% and 15%), and the technical-
4 Civil servants were asked to rate the importance of knowing a politician or someone with
political links for recruitment, promotions and pay rises on a scale of 1 to 7. As civil servants
may underreport the (sensitive) experience of having had political connections in the course of
their career in the public sector, we report in figure 1 the proportion of civil servants that
attaches at least some importance – i.e. not none at all – to political connections (scoring at
least 2 on the scale of 1-7).
Pg. 19 What works in civil service management
professional level (18%, 19% and 15%). Politicization thus permeates throughout
levels of hierarchies.
Figure 10. Proportion of civil servants who obtained their first job (also) thanks to
political connections, by level of hierarchy
Politicization also matters to some extent in all of the surveyed countries – albeit to a
greatly varying extent. As illustrated in figure 11, the share of civil servants for whom
political connections matter for recruitment, promotion and pay respectively vary from
5%-6% in Estonia to 39%-44% in Kosovo.5
5 In Chile, the indicator reports the proportion of civil servants who deem political connections
at least somewhat important for civil servants like them (rather than for the respondents’ own
recruitment, for instance).
Pg. 20 What works in civil service management
“Politicization of Figure 11. Proportion of civil servants who attribute at least some importance to
political connections for their recruitment, promotion or pay rises
civil service
management
practices
happens to
some extent in
all of the
surveyed
countries –
albeit to a
greatly varying
extent”
In sum, political connections matter for the recruitment, promotion and pay of a
significant – but varying – number of civil servants across levels of hierarchy and
countries.
1.2. How does politicization affect the behaviour and attitudes of civil servants?
How does civil service politicization affect civil servants? Prior studies suggest that the
politicization of recruitment has negative effects on civil service performance,
clientelism and corruption (Lewis 2008, Meyer-Sahling & Mikkelsen 2016, Oliveros &
Schuster 2017). Politicization can adversely affect the behaviour and attitudes of civil
servants through several channels (Meyer-Sahling, Mikkelsen & Schuster,
forthcoming). Most obviously, where political criteria have weight in personnel
decisions, competence is no longer (fully) prioritized, with negative consequences for
performance. Politicization also changes the career incentives of civil servants.
Responsiveness to political demands becomes the driver of career success – not
impartial service delivery to the public. Lastly, politicization can change the role
identity of civil servants. Civil servants become ‘political servants’ who are appointed
and promoted for their political service to politicians – not public service to society.
This shifts the sense of obligation of ‘public’ servants towards political superiors and,
potentially, away from the impartial and lawful exercise of their duties.
Our data confirms the negative effects of politicization, but provides a broader picture,
enabling us to assess the effects of the politicization of recruitment, promotion and pay
on a range of desirable attitudes and behaviour of civil servants across countries.
Controlling for a range of other factors6, we find that civil servants for whom political
connections were important for recruitment are less motivated to work hard, less
motivated to serve the public, less committed to staying in the public sector, less
performing and less satisfied with their jobs (figure 12). As illustrated in Appendix C
(figures c1 and c2), we find similar negative effects for the politicization of promotions
and pay rises.
6 For all regressions in this report, we run models which control for the gender, age, education,
level of hierarchy, income (in bands), years of experience in public sector, type of position
(dummy for contact with citizens) and country of the respondent. As noted above, the resulting
cross-sectional statistical associations are only suggestive of causal effects. Regression results
and the precise models (OLS, GLM or logit, for instance) used are available from the authors
upon request.
Pg. 22 What works in civil service management
“Politicization of Figure 12. The negative effects of politicized recruitment of civil servants
civil service
management
practices has
adverse effects
on work
motivation,
performance,
public service
motivation, job
satisfaction,
commitment to
staying in the
public sector
and integrity.”
The adverse effects of politicization do not stop at work and public service motivation,
satisfaction and performance, however. Our data suggests that they also extend to
integrity. As illustrated in figure 13, civil servants for whom knowing a politician or
someone with political links has been important for getting their job are more willing to
accept money or a personal present in exchange for helping someone through their
public sector position (controlling for the aforementioned factors). They have also
more frequently done so in the past, and are, in addition, more likely to help the
election campaign of a political party. In other words, politicization is associated with
greater corruption and clientelism in the public service (as measured by our list
experiments; see Appendix A.2). As illustrated in Appendix C, politicized promotions
and pay rises are, similarly, associated with greater clientelism. These latter results
are particularly intuitive: civil servants who owe their jobs to politicians are more likely
to help them in their (re-) election efforts.
Pg. 23 What works in civil service management
Figure 13. The effects of politicized recruitment on the integrity of civil servants
In sum, political connections not only matter in a significant share of civil service
management decisions. They also negatively affect the behaviour and attitudes of civil
servants, curbing work and public service motivation, job satisfaction and performance
while leading to more corruption and clientelism in public service. This puts a premium
on understanding which civil service management practices can depoliticize civil
services. This is assessed next.
As illustrated in figure 14, most public sector jobs in our surveyed civil services appear
to be advertised. Only a minority of civil servants (27%) has heard about their jobs
through word of mouth only (rather than some form of wider advertisement). This
share does reach 41% and 39% in Chile and Estonia respectively, however,
Pg. 24 What works in civil service management
“Some civil suggesting that a potential lack of advertisement precludes citizens in some countries
from applying to a significant minority of positions. Once applications are received,
servants have
interviews are the most common form of assessment in our surveyed countries (70%
obtained public of civil servants), followed by exams (49% of civil servants).
sector jobs
For both selection methods, however, there is significant variation across countries. In
which were some countries, the use of written examinations is highly common (Bangladesh, Nepal
neither and Brazil). In others, written exams are hardly ever used (Estonia and Ghana).
advertised nor Similarly, assessing candidates through interviews is highly common in most countries
assessed (Bangladesh, Malawi, Uganda, Kosovo, Ghana, Estonia and Nepal), but virtually
absent in Brazil (where, in most institutions, only managerial-rank civil servants are
through an
interviewed).
interview or
exam.”
Figure 14. Proportion of civil servants hired through formally merit-based recruitment
and selection procedures, by country
Written exams tend to coincide to some extent with job advertisements (r=0.22), while
the use of exams and interviews (r=-0.03) and job advertisements and interviews
(r=0.02) is largely unrelated. This reflects that, in the countries in our sample, most
civil servants apply for advertised jobs and are then assessed through written exams
or interviews (64%), yet fewer apply for advertised jobs and are then assessed
through both exams and interviews (28%). A relevant minority of civil servants also
Pg. 25 What works in civil service management
“Consistently entered the public sector through a job, which was not advertised and not assessed
through either an interview or an exam (6%), or advertised but not assessed through
advertising
an exam or interview (15%).
public sector
jobs and To what extent are these formal merit procedures effective in curbing politicization in
the civil service? Figure 15 suggests that formal procedures – advertisements of
assessing
public sector jobs, written examinations and interviews – can curb politicization. Both
candidates written examinations and interviews are associated with less importance of political
through written connections in recruitment. to the effect of job advertisements is substantively even
or oral larger, which underscores the importance of advertising job opportunities in the public
examinations sector – rather than merely disseminating them through word of mouth.
can reduce
politicization.” Figure 15. Effects of advertisements, written examinations and interviews on the
politicization of recruitment
Formal merit procedures are effective in curbing civil service politicization not only on
average, but also in almost all of the countries studied. As figure 16 illustrates, exams
exert a negative effect on politicization in 8 of the 10 countries studied (with one
negative effect not being significant); interviews exert a significant negative effect in 6
of the 10 countries studied (and are only statistically significantly positively associated
Pg. 26 What works in civil service management
with more politicization in one country); and job advertisements, similarly, are
associated with less politicization in 6 out of 10 countries (the two positive effects are
not significant).
In short, formal merit procedures are frequently – but far from always – used across
countries. On average and in most countries studied, they tend to curb politicization.
Particularly countries with high levels of politicization would thus benefit from
expanding formal merit safeguards – consistently advertising positions for recruitment
and promotion, and assessing civil servants through exams and/or interviews for
recruitment and promotion.
Pg. 27 What works in civil service management
7 Civil servants were asked to rate the importance of having friends, family or other
acquaintances in the public sector for having obtained their first job, for promotions and for pay
rises on a scale of 1 to 7. As civil servants may underreport the (sensitive) experience of
having had personal connections to get a job, promotions or pay increases, we report in figure
2 the proportion of civil servants that attaches at least some importance to personal
connections (scoring at least 2 on the scale of 1-7).
Pg. 29 What works in civil service management
Figure 17. Proportion of civil servants who obtained their first job (also) thanks to
personal connections inside the state, by level of hierarchy
Personal connections matter not only throughout hierarchies but also across countries
– but to a greatly varying extent. Personal connections matter in the recruitment of
most civil servants in Nepal (76%), for instance, but only a small minority of
(managerial) civil servants in Brazil (19%). Similarly, they matter for the promotion of a
majority (52%) of civil servants in Kosovo, yet for only a small minority (17%) in
Uganda; and for the pay rises of a large minority of civil servants in Kosovo (42%), yet
only for a minor share (5%) of civil servants in Bangladesh.
Pg. 30 What works in civil service management
Figure 18. Proportion of civil servants who attribute at least some importance to
“The incidence personal connections for their recruitment, promotion or pay rises8
of nepotism is
more
widespread in
our surveyed
civil services
than
politicization.”
8 In Chile, this question was only fielded in the Treasury, with an indicator which reports the
proportion of civil servants who deem personal connections at least somewhat important for
civil servants like them. As such, differences between Chile and the remaining countries maybe
due to differences in measurement.
Pg. 31 What works in civil service management
In sum, having support from friends, family members or other acquaintances inside
the state matters for the recruitment, promotion and pay rises of a significant minority
of civil servants across countries and levels of hierarchy. In fact, the incidence of
personal connections is more widespread than politicization. As noted at the outset,
this need not be bad news. Network-based personnel decisions need not have
adverse effects, but would do if they reflect personal favouritism and thus nepotism.
The next section thus assesses the effects of civil service ‘personalization’.
Pg. 32 What works in civil service management
Figure 20. The negative effects of personal connection-based civil service recruitment
Pg. 33 What works in civil service management
that personal connections are often, concurrently, political connections. Personal and
“Formally political networks of civil servants thus often appear to overlap, with both permeating a
meritocratic civil significant minority of personnel decisions.
service
In summary, nepotism affects a significant minority of civil service management
management
decisions in developing countries and is, in fact, more widespread than politicization. It
practices – such adversely affects several behaviour and attitudes of civil servants measured in our
as written exams survey: work and public service motivation, job satisfaction, performance, corruption
and interviews – and clientelism. This puts a premium on understanding which civil service
can curb management practices can curb nepotism. This is discussed in the next section.
The effects of exams and job advertisements generalize across almost all of the
countries studied (figure 23). At the country level, job advertisements exert negative
effects on nepotism in all of the countries studied, while exams exert negative effects
in 9 out of 10 countries. By contrast, the effects of interviews are somewhat mixed,
with interviews associated with less nepotism in the majority (but far from all)
countries.
In short, formal merit procedures and, in particular, written examinations for entry and
the advertisement of job positions tend to curb nepotism. Countries with high levels of
nepotism and/or politicization in civil service management would thus do well to
expand formal merit safeguards.
Pg. 35 What works in civil service management
Civil services vary significantly in their performance orientation: the extent to which
performance is perceived to matter for promotion, pay and dismissal prospects. We
find that performance-oriented civil service management has positive effects. Civil
servants are more satisfied, committed and, at times, motivated to serve the public,
work hard and perform where they perceive that performance shapes their pay,
promotion and job stability prospects. Making sure that performance matters in civil
service management decisions thus brings important payoffs. Formal civil service
management practices can play a role. Performance evaluations are positively
associated with (some) perceptions of performance orientation. Their effects
depend, however. Where performance objectives are not identified beforehand and
evaluation results are not perceived to matter for career advancement, pay and
promotion prospects, they have counterproductive effects. The evidence thus
underscores the importance of ensuring both that performance matters in civil
service management decisions, and that formal performance management systems
are designed and implemented well to achieve this end.
The role of performance orientation in civil service management has been discussed
extensively in debates surrounding the introduction of New Public Management
reforms in developing countries (Schick 1998, Manning 2001). Proponents of New
Public Management reforms point to the importance of rewarding performance as an
incentive for better performance and motivation in the civil service. Evaluations that
have focused on performance-related pay have provided qualified support for this
perspective – albeit with hardly any studies assessing the core civil service (Hasnain
et al. 2012).
Our survey provides evidence which can (to some extent) help adjudicate between
“Contrary to these perspectives. We thereby shed light on the effects of performance-oriented civil
popular service management. We understand performance orientation as ensuring that work
stereotypes performance matters in pay, promotion, transfer and dismissal decisions. This, by no
about means, requires a mechanistic pay-for-performance or promotion scheme in which
performance ratings automatically translate into pay rises or promotions (see OECD,
bureaucracies,
2005). Rather, a performance orientation provides civil servants with a sense that
there is some performance is (eventually) worthwhile – because career, pay or job stability prospects
sense of improve.9
performance Figure 24 illustrates that, contrary to popular stereotypes about bureaucracies, there is
orientation in all some sense of performance orientation in all of the surveyed civil services. Civil
of the surveyed servants on average in our sample, for instance, either ‘neither agree nor disagree’ or
civil services.” ‘somewhat agree’ that they could be dismissed for poor performance. Cross-country
variation is significant, however. In Estonia, Chile, Kosovo and Albania, civil servants,
on average, somewhat agree that they could be dismissed for poor performance. At
the other extreme, in Bangladesh, the mean response is between ‘somewhat
disagree’ and ‘neither agree nor disagree’. Similar variation can be observed for the
performance orientation of promotions. On average, civil servants rate the importance
of work performance for their career advancement at 3.2 on a scale of 0 to 4. Only 7%
attribute no importance whatsoever to their performance. This suggests that the large
majority of civil servants, in fact, perceive their performance to matter at least in some
way for their career advancement. The share of civil servants for whom this is not the
case varies significantly across countries, however, from 2% in Estonia and Ghana to
25% in Bangladesh. Lastly, across most countries, there is scepticism about the
performance orientation of pay decisions. With the notable exception of Estonia, civil
servants in all countries are more likely to disagree than agree that their work
performance has influenced their pay (mean of 1.4 on a scale of 0 to 4, where 4 is
strongly agree).
9 We measure performance orientation in pay with the extent of agreement to the statement
‘My work performance has had an influence on my salary in the public service.’ Performance
orientation in dismissals with the statement ‘I might be dismissed from the civil service if I do
not perform well.’ And performance orientation in promotion by asking respondents to rate on a
scale from 1 to 7 how important work performance is for their future career advancement. We
normalized the 1-7 scale to a 0-4 scale in figure 15.
Pg. 38 What works in civil service management
Figure 24. Civil servants’ perceived performance for promotion prospects, pay rises
and dismissals
Figure 24 thus plausibly suggests that there is some performance orientation across
countries, albeit much more so in some countries (e.g. Estonia) than others (e.g.
Bangladesh); and that performance orientation is relatively less pronounced in pay
setting (with the caveat that some of these differences may stem from different
variable measurements). Moreover, the figure suggests that performance orientation
in one civil service management function need not coincide with performance
orientation in another. In fact, performance orientations in promotions and salaries
(r=0.15), promotions and dismissals (r=0.13) and pay and dismissals (r=0.2) are only
weakly correlated. Introducing a performance orientation in one civil service
management function thus does not seem to spill over into performance orientations in
other civil service management functions. The next section will assess the extent to
which such performance reforms matter.
Having the perception that performance matters for their career advancement, pay
and job stability enhances civil servants’ job satisfaction and preference for remaining
Pg. 40 What works in civil service management
in the public sector. Performance-oriented promotions and job stability (albeit not pay)
“Contrary to also have a small positive effect on public service motivation; and performance-
popular oriented promotions and pay (albeit not dismissals) positively affect work motivation
stereotypes, civil and performance.
servants are At the same time, we observe hardly any effects of performance orientation on
more satisfied integrity. As illustrated in figure 26, performance-oriented promotions have no
where civil significant effects on any of our integrity measures, and neither do performance-
service oriented dismissals (Figure c12 in the Appendix C). Performance-oriented pay is
statistically significantly associated with less nepotism – albeit none of the other four
management is integrity measures. This might be a statistical artefact, however (of finding, by random
performance- chance, an effect in one in fifteen regressions), and thus only provides tentative
oriented.” evidence for a positive effect of performance orientation on integrity.
What can be learned from these findings? First of all, they suggest that, contrary to
popular stereotypes, civil servants appreciate some performance orientation in civil
service management decisions. The most robust effects of performance orientation
across the three civil service management functions are on job satisfaction and
preference for working in the public sector. Unintendedly, giving civil servants a sense
Pg. 41 What works in civil service management
that their performance matters may thus affect job satisfaction and retention more than
“Basing performance or work motivation.
promotions on
Second, a performance orientation in and of itself does not appear to crowd out public
performance
service motivation (PSM); in fact, we observe ‘crowding in’ of PSM for performance-
might be more oriented promotions and dismissals based on our cross-sectional data. At the same
important than time – and contrary to the lessons about nepotism and politicization – a performance
performance- orientation only appears to have a, at best, tentative positive effect on integrity.
related pay or Further analyses would be required to assess whether it might shape integrity
indirectly, however (e.g. by fostering PSM).
dismissals.”
Fourth, performance-oriented promotions might trump performance-oriented pay and
dismissals in making civil services work. New Public Management reforms have
frequently focused on pay-for-performance and temporary contracts as ‘hard’
performance incentives. In our data, by contrast, only performance-oriented
promotions are significantly associated with performance, work motivation,
performance, satisfaction and a preference for public sector jobs. This suggests that
reformers would do well to consider performance orientation in career advancement in
their reform designs. Lastly and most obviously, the positive effects of performance
orientation put a premium on understanding which civil service management practices
can bring this orientation about. This is analysed next.
How can reformers and civil service managers increase the perception among civil
servants that performance matters for their promotion, pay and job stability? In this
section, we assess the effectiveness of what has arguably been the core instrument to
this end in managerial reforms: performance evaluation systems that regularly assess
the work performance of civil servants. Performance evaluation systems, of course,
come in many guises: they differ in the frequency of evaluations, whether objectives
are set in advance, whether targets are quantitative or qualitative, whether results of
evaluations are discussed, who evaluates, and whether there is a forced distribution of
ratings among staff in a department, among many.
Our survey is not able to capture the manifold dimensions of performance evaluation
systems. Rather, our ambition is more modest: the survey can shed light on whether
having evaluations has any effects on performance orientation; and whether two basic
design features of performance evaluations – setting goals before an evaluation
Pg. 42 What works in civil service management
period and linking evaluation results to promotion, pay or job stability prospects –
shape the effects that performance evaluations have on civil servants.
Figure 29. The effects of having had a performance evaluation on the perceived
performance orientation of civil service management, by country
agreeing on objectives prior to an evaluation period (to ensure civil servants know
what to perform towards); and (2) giving civil servants a sense that evaluation results
matter (by affecting their prospects for promotion, pay and/or job stability in some
way). We find that countries vary in the extent to which these basic design features
are in place; and that evaluations can have positive or negative effects depending on
their design.
Figures 31 and 32 illustrate that this variation in design matters for the effects of
“Whether performance evaluations. Performance evaluations are positively associated with the
performance perceived performance orientation of pay and promotions where they are based on
evaluations have objectives that were agreed before the assessment period (the effect on the
positive or performance orientation of dismissals is not significant). In contrast, performance
evaluations have negative effects on the perceived performance orientation of
negative effects
dismissals and no significant effects on the perceived performance orientation of pay
depends in part where no prior performance objectives were agreed (the effect on performance-
on whether oriented promotions remains positive). This provides suggestive evidence that the
evaluation effect of performance evaluations is in part contingent on whether performance
objectives are set in advance. Where they are not, they may be counterproductive.
objectives are
This finding may not surprise: being evaluated against objectives which are set ex
set in advance. post is likely seen as arbitrary and unfair by civil servants.
We find similar results for a second core design feature: the link between performance
evaluation results and pay, promotion and dismissal prospects. Where civil servants
perceive such links, the effects of performance evaluations on perceptions of
performance orientation of pay, promotions and dismissals are invariably positive. By
Pg. 47 What works in civil service management
contrast, where civil servants do not see a link between performance evaluations
“Performance results and decisions about pay, promotions and dismissals, the effects of
evaluations only performance evaluations on the perceived performance orientation of pay and
have positive dismissals are negative (and the effect on performance-oriented promotions is
effects where insignificant). This suggests, once again, that performance system designs shape
whether they make civil services work or achieve the opposite.
they are
designed and
implemented Figure 32. The effects of performance evaluations on the perceived performance
well. Where this orientation of civil service management depend on whether civil servants believe that
evaluation results affect dismissals, pay and promotion
is not the case,
they have
negative effects
on civil servants.
The implication for civil service reformers is clear: make sure that performance matters
in civil service management decisions. How can reformers do so? We find that
Pg. 48 What works in civil service management
performance evaluations are one important tool. Their effects, however, crucially
depend on design. Where performance objectives are identified beforehand and
evaluation results matter in the eyes of civil servants for their career advancement,
pay and promotion prospects, they tend to have positive effects on perceptions of
performance orientations. Without these basic design features in place, they can be
counterproductive. This underscores the importance of good design in performance
management systems to achieve positive – rather than negative – effects on
performance orientation in the civil service.
Pg. 49 What works in civil service management
How do civil servants in developing countries perceive their pay? Taking a ten-
country average, they are relatively dissatisfied with their pay and do not find it
sufficient to maintain their households. At the same time, most would find it hard to
find a better-paid private sector job. Salaries are thus not necessarily uncompetitive,
even if they are perceived as unsatisfactory and insufficient. Notwithstanding these
averages, pay perceptions vary significantly across countries, institutions and
groups of civil servants. This variation matters. Higher pay satisfaction and
sufficiency are associated with greater job satisfaction and intent to remain in the
public sector. We do not find significant positive effects on work motivation and
performance, however. At the same time, pay satisfaction and sufficiency may affect
motivation and performance indirectly: by discouraging departures of more
motivated and performing staff, who also deem it easier to find better-paid private
sector jobs. As higher pay is a significant driver of greater pay satisfaction and
sufficiency, our fourth lesson is clear: pay enough to retain the (more) motivated and
performing staff. Where retention (or attraction) of motivated and performing staff is
not a challenge, however, higher pay may do little to enhance motivation and
performance of staff.
Pay can affect civil servants in a myriad of ways. Advocates of higher pay in public
sectors point to studies linking pay increases to the attraction of higher performing
candidates to public service (see, e.g., Nickell & Quintini, 2002; Dal Bo, Finan & Rossi,
2013); to greater work motivation and job satisfaction as well as reduced turnover
once civil servants are hired (see, e.g., Esteve et al. 2017; Grissom, Viano & Selin,
2015); and to reduced corruption, as civil servants are less likely to need to engage in
corruption to make a living for their family and face greater opportunity costs when
caught (see, e.g. van Rijckeghem & Weder 2001).
Others have contested these arguments. Herzberg (1987) classically posited that high
pay cannot motivate staff; low pay, however, can dissatisfy (and thus demotivate)
staff. Some have gone further and argued for adverse consequences of high pay
(beyond their fiscal cost). High pay is, for instance, feared to crowd-out the intrinsic
and public service motivation of civil servants, due to the attraction of less public-
Pg. 50 What works in civil service management
spirited employees and the greater focus on material benefits of public sector work
“Most civil (see, e.g., Navot et al. 2015). Some studies have found empirical support for this
servants are notion, identifying, for instance, positive associations between pay levels and
dissatisfied with corruption (e.g. Karahan et al. 2006, Navot et al 2015); or, at least, no significant
their salaries association (e.g. Dahlstroem et al 2012, Rauch & Evans 2000, Treisman 2000).
and find them
The number of studies which have assessed these arguments with data from civil
insufficient to servants in developing countries, however, remains limited. Which (if any) of these
maintain their arguments robustly apply to developing country civil services thus remains uncertain.
households. Yet, With our data, we can shed light on this.
only a minority
Departing from most prior studies, we will, in a first step, assess the perceptions of
would find it pay of civil servants. The rationale is simple: many of the theoretical mechanisms
easy to find a linking pay to behaviour ultimately rest not on pay itself, but on the perceptions that
better-paid job civil servants have of their pay. Whether pay acts as a dissatisfier in Herzberg’s theory
outside the depends first and foremost on how civil servants perceive their pay, for instance.
Perceptions of pay are, of course, driven by pay. As we will show below, they also
public sector.” depend on other factors, however. This puts a premium on assessing pay perceptions.
In our survey, we captured three important perceptions of pay: whether civil servants
are satisfied with their pay (salary satisfaction); whether they could sustain their
household through their salary alone (salary sufficiency); and how easy it would be for
them to find a better-paid job outside the public sector (as a proxy for salary
competitiveness). The three items were measured on a five-point Likert scale, ranging
from strongly disagree to strongly agree.
On average, civil servants in our sample are relatively dissatisfied with their salaries
and do not find them sufficient – but, similarly, would not find it easy to find a better
paid job in the private sector. As illustrated in figure 33, merely 37% of our surveyed
civil servants somewhat agree or strongly agree that they are satisfied with their
salary; and 40% that their salary alone is sufficient to sustain their household. Pay
satisfaction and sufficiency are thereby relatively closely related (r = 0.54). At the
same time, only 39% strongly or somewhat agree that it would be easy for them find a
better-paid private sector job. This is congruent with studies showing that most
developing country public sectors feature wage premiums relative to the private sector
(see, e.g. Finan & Pande, 2017). Remarkably, perceived salary competitiveness and
pay satisfaction (r = 0.16) and sufficiency (r = 0.10) are thereby only weakly linked.
The judgments of civil servants about their salaries (in terms of satisfaction and
Pg. 51 What works in civil service management
sufficiency) are thus far from always linked to the competitiveness of their salaries –
an insight often overlooked in pay reforms to enhance motivation and performance.
In summary, the data thus suggests that salaries are perceived as at least somewhat
competitive, but nonetheless not necessarily as satisfactory or sufficient; and that
perceptions of satisfaction and sufficiency are often delinked from salary
competitiveness.
In other words, in a subset of developing country civil services, civil servants are
overwhelmingly dissatisfied with their salaries, and find them insufficient to maintain
their households. Again, this need not mean that salaries are uncompetitive, however.
Pg. 52 What works in civil service management
Even in the countries with the least salary satisfaction (Malawi, Ghana and Uganda),
between 41% and 53% of civil servants would not find it easy to find a better paid
private-sector job; yet only 9%-16% are satisfied with their pay.
Perceptions of pay vary not only across countries, but also within countries across
institutions and groups of civil servants. To illustrate, figure 34 compares pay
perceptions by rank in hierarchy. Most prior studies have shown larger public sector
pay premiums (over private sector pay in comparable positions) at lower levels in the
hierarchy (see, e.g. Panizza, 2001). Our data is congruent with – but nuances – these
studies. Relative to managers (54%), more technical-professional (63%) and
administrative support (59%) staff would find it difficult to find a better paid private-
sector job. At the same time, more managers than professional-technical and
administrative support staff are satisfied with their (higher) pay (46% relative to 39%
and 28%) and find it sufficient to maintain their households (47% relative to 43% and
27%). Managers are thus relatively (more) satisfied with their pay despite it being
relatively less competitive.
This finding provides further suggestive evidence that the effects of pay cannot be
“The effects of gleaned from solely looking at pay levels. The differential perceptions across
pay depend not hierarchies also provide additional evidence for the need to tailor reforms – in this
only on pay case pay reforms – to the realities – in this case pay perceptions – of different groups
levels, but also and institutions within the civil service. This, of course, presupposes that pay
perceptions matter for the attitudes and behaviour of civil servants. That this is, in fact,
on the
the case is evidenced next.
perceptions civil
servants have of 4.2. How do pay perceptions affect the attitudes and behaviour of civil servants?
their pay.”
As noted in the previous section, prior studies offer competing arguments and findings
about the effects of pay. Some associate higher pay with greater job satisfaction, work
motivation, public sector commitment and integrity. Others find no significant – or
negative – effects. Our cross-sectional data can shed light on these effects in
developing country civil service contexts. One note of caution about the limits of our
data is due, however: pay can as much determine attitudes and behaviour (such as
work motivation and performance) as be determined by them. Our cross-sectional
data does not enable us to fully disentangle reverse causality – which, for other
variables such as those related to initial recruitment in lessons 1 and 2, is a somewhat
lesser concern. As such, our data only provides highly suggestive evidence on the
effects of pay.
With this caveat in mind, we find that higher pay satisfaction and sufficiency are
strongly positively associated with job satisfaction and a preference for remaining in
the public sector. However, we do not find significant positive associations with work
motivation and performance (and only a small positive association between salary
satisfaction and public service motivation).10 Pay satisfaction and sufficiency might
thus matter more for retaining (satisfied) civil servants and reducing public sector
turnover, than for motivating hard work or performance.
10In fact, the effect of salary satisfaction on performance is negative. As illustrated in Appendix
c.13, however, this is likely due to reverse causality: more performing civil servants are less
satisfied with their salaries.
Pg. 54 What works in civil service management
satisfaction and a preference for remaining in the public sector. This need not
“Higher pay may surprise: those who find it harder to get a better-paid private sector job are more likely
plausibly affect to want to stay in the public sector.
work motivation
and However, greater salary competitiveness is also associated with less work motivation
and performance. This, of course, need not mean that competitive salaries reduce
performance
work motivation and performance. Rather, as illustrated in Appendix D, reverse
indirectly: by causality is likely at play. More motivated and performing civil servants are also those
helping retain who find it easier to find a private sector job that pays better. As those who find it
more motivated easier to find private sector jobs that pay better also have a greater preference for
leaving the public sector, salary competitiveness may plausibly affect work motivation
staff who can
and performance in public sectors indirectly: by helping retain more motivated and
more easily find performing staff members who deem it easier to find a better-paid private sector job.
better-paid
private sector Figure 37. The effects of (perceived) salary competitiveness on civil servants
jobs.”
The effects of pay perceptions on integrity are more subdued. We find no significant
effects of pay satisfaction and competitiveness on any of our integrity measures
(Appendix D, figures d2 and d3); but one significant (and theoretically intuitive) effect
of salary sufficiency. In line with needs-based corruption accounts (see, e.g. Bauhr
2017), civil servants who deem their salaries less sufficient to maintain their
households are also more likely to indicate that they have used their public sector
position to benefit family members, friends or other personal acquaintances (in a list
Pg. 56 What works in civil service management
experiment; see Appendix A.2). As we only identify a single (albeit highly plausible)
effect in one in fifteen integrity regressions, however, we cannot rule out that this
finding merely occurred by statistical chance.
In summary, our data indicates suggestively that pay can play an important direct role
in job satisfaction and staff retention, and an indirect role in staff motivation and
performance by helping retain more motivated and performing types who can more
easily find better-paid private sector jobs. Potentially, higher may also curb needs-
based integrity violations. These inferences, of course, assume that pay, in fact, drives
the perceptions of pay we assessed in this section. This is explored next.
As noted above, pay perceptions ultimately shape how pay affects the attitudes and
behaviour of civil servants. This puts a premium on understanding how institutions can
foster favourable pay perceptions. Higher pay is, of course, the most obvious tool. In
fact, as illustrated in figure 39, higher pay is associated with both higher pay
satisfaction and higher perceived pay sufficiency, with almost every additional income
band boosting pay satisfaction and sufficiency (controlling for our usual variables). The
effect sizes are, moreover, substantively large. Going from the lowest to the highest
income band lifts pay satisfaction by 1.8 (on a 0-4 scale), for instance. The resulting
Pg. 57 What works in civil service management
conclusion does not need to surprise: higher pay can enhance perceptions of pay
“Higher pay can satisfaction and sufficiency.
improve pay
satisfaction and Figure 39. Effects of higher income and interesting work on pay satisfaction and
(perceived) sufficiency (relative to lowest salary band)
sufficiency, but
is not the only
management
practice shaping
pay
perceptions.”
In summary, higher pay can enhance pay satisfaction and sufficiency, which in turn
helps retain staff and enhance job satisfaction. Indirectly, it may also foster work
motivation in public sectors by avoiding departures of more motivated and performing
Pg. 58 What works in civil service management
staff. Our final lesson is thus clear: pay civil servants enough to retain more motivated
and performing staff. This complements prior research which had pointed to the
importance of sufficient pay to attract motivated and able staff (Dal Bo, Finan & Rossi,
2013).
How much is ‘enough’ will, of course, vary. As noted, pay perceptions and turnover
intent vary across countries, institutions and groups of civil servants. Where turnover
intent is low or pay satisfaction, sufficiency or competitiveness are high, higher pay
may do little for the motivation and performance of (existing) staff. At the same time,
where only few motivated staff could be retained by large pay increases for all civil
servants (in cases of collective bargaining, for instance), higher pay may not be cost-
effective. Where significant turnover of high performing staff can be forestalled by
marginal pay increases, however, seeking pay rises may be well warranted.
Pg. 59 Implications for civil service reform
First of all, that civil service reforms ultimately require tailoring to the realities of each
institution – and, at times, the realities of each unit or group of professionals within
institutions. Country-level civil service reform programs would do well to keep this in
mind. Adapting institutional reforms to local realities has, of course, become a mantra
in governance reform discourse (Andrews, 2013). As such, this implication might
seem like old wine in new bottles. Our survey can shed more specific light on what
‘adapting to local realities’ means for civil service reform, however.
As noted in chapter 2, the attitudes and behaviour of civil servants vary sharply across
and within countries and institutions. The resulting pattern bears little resemblance to
conventional wisdoms about islands of excellence in seas of mediocrity in developing
country states. Instead, gradual differences rather than dichotomies between poor and
strong performance mark most institutions. Moreover, institutions (and civil servants)
which score highly in one attitude or behaviour (e.g. work motivation) often do not do
so in another (e.g. commitment of civil servants to remain in public sector). Institutions
may thus have strengths in some dimensions of civil servant behaviour and attitudes,
while having weaknesses in others. Designing effective civil service reforms thus
requires, as a first step, an understanding of the attitudes and behaviour of the civil
servants (for instance via staff survey) to be able to shape them for the better. In light
of the diversity of these attitudes and behaviour across and within institutions, cookie
cutter civil service reforms cannot be effective. Rather, they require adaptation to the
realities of each institution or group of civil servants within institutions; and those
realities comprise both an understanding of existing civil service management
practices, and of the attitudes and behaviour of civil servants that are being shaped by
them.
Once these local realities are understood, reforms can be tailored to them. Of course,
a panoply of reforms might be effective in any given context. Notwithstanding, our
survey identified four reforms which had positive effects in most of the countries we
Pg. 60 Implications for civil service reform
studied. In other words, they tended to lead to more motivated, committed, satisfied,
performing and ethical civil servants.
For long-time civil service reformers, these lessons may not come as a surprise.
Advocates of Weberian bureaucracies have long argued for meritocratic personnel
practices to curb politicization and nepotism; managerial reform proponents in turn
have long advocated performance management systems to incentivize hard work.
To some extent, this is good news: it suggests that many prior reform prescriptions
have the potential to improve civil services. It is also bad news, however. Institutions
frequently do not adopt practices that are statistically associated with more motivated,
committed and ethical civil servants. This, of course, points to the need to understand
civil service reform not only as a technical design, but also as a political and
implementation challenge. Our report principally sought to inform reform design. With
that said, we hope that the evidence presented on reform effects can also help
reformers make their case and convince others of the benefits and importance of civil
service reform and effective implementation.
Pg. 61 Appendix
Appendix
Appendix A.1: civil service surveys
11The 14 institutions surveyed in Brazil were: Ministry of Finance, Treasury, Tax Administration
(RFB), Ministry of Planning, Social Security Institute (INSS), Ministry of Industry, Ministry of
Culture, General Audit Institution, Ministry of Transport, Ministry of Social Development,
Ministry for Urban Affairs, Ministry of Environment, Ministry of Science and Technology and the
Attorney General. The survey focused on civil servants on Brasilia where location data was
available.
12 The 11 institutions surveyed in Chile were: Treasury, Social Security Institute (IPS),
Economic Development Agency (CORFO), Civil Service Agency, National Health Fund
(FONASA), Department of Planning in the Ministry of Public Works, Medical Legal Service, the
Department for Archives, Libraries and Museums and the National Fisheries and Aquaculture
Service.
Pg. 62 Appendix
On a scale of 1 to 7, how satisfied or dissatisfied are you with your job? (Scale: 1:
Completely Dissatisfied; 7 Completely Satisfied)
How frequently do the following statements apply to you? (Scale: from ‘Never’ to
‘Always or almost Always’)
I put forth my best effort to get my job done regardless of any difficulties
Self-assessed performance
How frequently do the following statements apply to you? (Scale: from ‘Never’ to
‘Always or almost Always’)
Imagine that, hypothetically speaking, you had to find a new job in the next few
months, in which sector would you prefer to search for a job? (Response options:
Public sector // Private sector)
To what extent do you agree with the following statements? (Scale: ‘Strongly
Disagree’ to ‘Strongly Agree’)
It is important that citizens can rely on the continuous provision of public services
It is fundamental that the interests of future generations are taken into account when
developing public policies
I get very upset when I see other people being treated unfairly
I would agree to a good plan to make a better life for the poor, even if it costs
me money
Public servants sometimes receive offers in the course of their work life or daily
activities. Below is a list of several hypothetical offers. Please indicate how many (not
which) of them you would accept.
If I were offered a public job in a region I do not know well, I would accept it.
If I were offered the opportunity to teach classes at university about my field of work, I
would accept it.
If I were offered a better paid job in exchange for taking on broader responsibilities at
work, I would accept it.
There are many more activities that public servants undertake in the course of their
jobs and daily lives. Below are four lists with examples of such activities. For each list,
please indicate how many of the listed activities you have undertaken. Do not tell the
interviewer or anyone else which you have undertaken.
In regards to requests and opportunities in your job, how many of the following
activities have you undertaken in the past two years.
You have received a gift or otherwise personally benefited from your position.
You have accepted a request from your manager to represent your institution at a
public media event.
Integrity: nepotism
In regards to the relationship between your work and family and friends, how many of
the following activities have you undertaken in the past two years.
You have considered leaving your job to spend more time with your family.
You have discussed with friends or family how reconcile work obligations with
obligations at home.
You have discussed with a friend or a family member the advantages and
disadvantages of working in the public sector.
You have used your position to help members of your family or friends.
Integrity: clientelism
In regards to elections, how many of the following activities have you undertaken
during the last national campaign.
You voted.
You helped divert government resources to a party or person with political links.
You helped your manager with an important assignment for your organisation.
Figure c2. The negative effects of politicized pay rises in the civil service
Pg. 70 Appendix
Figure c3. The effects of politicized promotions on integrity and impartiality in the civil
service
Figure c4. The effects of politicized pay rises on integrity and impartiality in the civil
service
Pg. 71 Appendix
Figure c13. Reverse causality checks: effects of performance and work motivation on
salary satisfaction and competitiveness, and of salary satisfaction on perceived salary
competitiveness
Pg. 76 Appendix
Bibliography
Adams, J.S. (1965) “Inequity in Social Exchange”. In L. Berkowitz (Ed.), Advances in
Experimental Social Psychology, Vol. 2, pp. 267-299. New York: Academic Press.
Brown, S., Gray, D., McHardy, J. and Taylor, K. (2015) “Employee Trust and
Workplace Performance”. Journal of Economic Behavior & Organization. Vol. 116, p.
361-378
Chen, C.-A., and Hsieh. C.-W. (2015) “Does Pursuing External Incentives
Compromise Public Service Motivation? Comparing the Effects of Job Security and
High Pay”. Public Management Review, Vol. 17(8), p. 1190-1213.
Dal Bó, E., Finan, F., and Rossi, M. (2013) “Strengthening State Capabilities: The
Role of Financial Incentives in the Call to Public Service”. The Quarterly Journal of
Economics, Vol. 128(3), p. 1169–1218.
Dahlström, C., Lapuente, V., and Teorell, J. (2012) “The Merit of Meritocratization:
Politics, Bureaucracy, and the Institutional Deterrents of Corruption.” Political
Research Quarterly, Vol. 65(3), p. 656–68.
Dineva, B., Holbrook, L., and Geshuri, A. (2015) “Data: Referrals Strongly Impact
Retention and Depend on Employee Performance”. LinkedIn Talent Blog, Available at
https://business.linkedin.com/talent-solutions/blog/2015/05/data-referrals-strongly-
impact-retention-and-depend-on-employee-performance [Accessed 05 February,
2018].
Esteve, M., Schuster, C., Albareda, A., and Losada, C. (2017) “The Effects of Doing
More with Less in the Public Sector: Evidence from a Large-Scale Survey”. Public
Administration Review, Vol. 77(4), p. 544–553.
Finan, F., and Pande, R. (2017) “The Personnel Economics of the Developing State”.
In A. Banerjee and E. Duflo (eds) Handbook of Field Experiments, Volume II.
Amsterdam: North Holland.
Grissom, J., Viano, S., and Selin, J. (2016) “Understanding Employee Turnover in the
Public Sector: Insights from Research on Teacher Mobility”. Public Administration
Review, Vol. 76(2), p. 241–251.
Hasnain, Z., Manning, N., and Pierskalla, J.-H., (2012) “Performance-related Pay in
the Public Sector A Review of Theory and Evidence”. Policy Research Working Paper
6043. Washington: World Bank Publications.
Pg. 79 Bibliography
Herzberg, F. (1987) “One More Time: How Do You Motivate Employees?”. Harvard
Business Review, September-October, p. 5-15.
Karahan, G.R., Razzolini, L., and Shughart, W.F. (2006) “No Pretense to Honesty:
County Government Corruption in Mississippi.” Economics of Governance, Vol. 7(3),
p. 211–27.
Kim, S., Vandenabeele, W., Wright, B.E., Bøgh Andersen, L., Cerase, F.P.,
Christensen, R.K., Desmarais, C., Koumenta, M., Leisink, P., Liu, B., Palidauskaite, J.,
Pedersen, L.H., Perry, J.L., Ritz, A., Taylor, J. and De Vivo, P. (2013) “Investigating
the Structure and Meaning of Public Service Motivation across Populations:
Developing an International Instrument and Addressing Issues of Measurement
Invariance”. Journal of Public Administration Research and Theory, Vol. 23(1), p. 79-
102.
Kopecký, P., Meyer Sahling, J.-H., Panizza, F., Scherlis, G., Schuster, C. and Spirova,
M. (2016) “Party Patronage in Contemporary Democracies: Results from an Expert
Survey in 22 Countries from Five Regions”. European Journal of Political Research,
Vol. 55(2), p. 416–431.
Meyer-Sahling, J.-H., and Mikkelsen, K.S. (2016) “Civil Service Laws, Merit,
Politicization, and Corruption: The Perspective of Public Officials from Five East
European Countries”. Public Administration, Vol. 94(4): 1105–1123.
Pg. 80 Bibliography
Navot, D., Reingewertz, Y., and Cohen, N. (2015) “Speed or Greed? High Wages and
Corruption Among Public Servants.” Administration and Society, Vol. 48(5), p. 580-
601.
Nickell, S. and Quintini, G. (2002) “The Consequences of the Decline in Public Sector
Pay in Britain: A Little bit of Evidence”. The Economic Journal, Vol. 112(447), p. F107–
F118.
Oliveros, V., and Schuster, C. (2017) “Merit, Tenure and Bureaucratic Behavior:
Evidence from a Conjoint Experiment in the Dominican Republic.” Comparative
Political Studies, early view.
Panizza, U. (2001) “Public Sector Wages and Bureaucratic Quality: Evidence from
Latin America”. Economía, Vol. 2(1), p. 97-139.
Ritz, A., Brewer, G. A., and Neumann, O. (2016) “Public Service Motivation: A
Systematic Literature Review and Outlook”. Public Administration Review, Vol. 76(3),
p. 414–426.
Pg. 81 Bibliography
Roggers, D. (2017) Who Serves the Poor? Surveying Civil Servants in the Developing
World. Policy Research Working Paper 8051. Development Research Group, Impact
Evaluation Team.
Schuster, C., Meyer-Sahling, J.-H., Mikkelsen, K.S., and Gonzáles Parrao, C. (2017)
Prácticas de Gestión de Personas para un Servicio Público más Motivado,
Comprometido y Ético en Chile. Evidencia de una Encuesta con 20.000 Servidores
Públicos en Chile y Otros Países. Informe preparado para la Dirección Nacional del
Servicio Civil, octubre.
Schick, A. (1998) “Why Most Developing Countries Should Not Try New Zealand
Reforms”. The World Bank Research Observer, Vol. 13(1), p. 123-131.
Van Rijckeghem, C. and Weder, B. (2001) “Bureaucratic Corruption and the Rate of
Temptation: Do Wages in the Civil Service Affect Corruption, and by How Much?”
Journal of Development Economics, Vol. 66(2), p. 307–31.
World Bank (2008) Public Sector Reform: What Works and Why? – An IEG Evaluation
of World Bank Support. Washington DC: The World Bank
Pg. 82 Bibliography
Wright, B. (2001) “Public Sector Work Motivation: A Review of the Literature and a
Revised Conceptual Model”. Journal of Public Administration Research and Theory,
Vol. 11(4), p. 559-586.