Policy Research Working Paper 10333 The Global Survey of Public Servants A Foundation for Research on Public Servants around the World Christian Schuster Kim Sass Mikkelsen Daniel Rogger Francis Fukuyama Zahid Hasnain Dinsha Mistree Jan Meyer-Sahling Katherine Bersch Kerenssa Kay Development Impact Evaluation Group & Governance Global Practice March 2023 Policy Research Working Paper 10333 Abstract How do civil service management practices differ within comparative public administration and the state differently, and across governments? How do core attitudes of public based on micro-data from actors who experience govern- servants—such as their motivation or satisfaction—differ ment first-hand. This paper introduces the Global Survey of within and across governments? Understanding how public Public Servants, a global initiative to collect and harmonize administrations around the world function and differ is large-scale, comparable survey data on public servants. The crucial for strengthening their effectiveness. Most compar- Global Survey of Public Servants can help scholars compare ative measures of bureaucracy rely on surveys of experts, public administrations around the world and understand households, or firms, rather than directly questioning the internal dynamics of governments, with the published bureaucrats. Direct surveys of public officials enable gov- Global Survey of Public Servants data freely available online. ernments to benchmark themselves and scholars to study This paper is a product of the Development Impact Evaluation Group, Development Economics and the Governance Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http:// www.worldbank.org/prwp. The authors may be contacted at drogger@worldbank.org and zhasnain@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team The Global Survey of Public Servants: A Foundation for Research on Public Servants around the World1 Christian Schuster ⓡ Kim Sass Mikkelsen ⓡ Daniel Rogger ⓡ Francis Fukuyama ⓡ Zahid Hasnain ⓡ Dinsha Mistree ⓡ Jan Meyer-Sahling ⓡ Katherine Bersch ⓡ Kerenssa Kay2 JEL classifications: C83, D73, H7, H83 Keywords: Survey Methods, Bureaucracy, Administrative Processes in Public Organizations, Corruption, Public Administration, State and Local Government 1 Certified random author order. The authors gratefully acknowledge excellent research assistance from Robert Lipinski, Ayesha Khurshid, and Srishti Gupta, as well as financial support for the underlying data collection from the World Bank’s i2i and DIME initiatives and Governance Global Practice, UKAID, the European Commission, the University College London (UCL) Global Engagement Fund, Nottingham and Stanford. Further contributors to the Global Survey of Public Servants are listed at https://www.globalsurveyofpublicservants.org/surveyteam. 2 Schuster: University College London; Mikkelsen: Roskilde University; Rogger, Kay, Hasnain: World Bank; Fukuyama, Mistree: Stanford University; Meyer-Sahling: University of Nottingham; Bersch: Davidson College. Introduction: Measurement of Public Administration and the Global Survey of Public Servants How governments are administered is a critical determinant of socio-economic outcomes (Evans and Rauch 1999; Finan, Olken and Pande 2015; Pepinsky, Pierskalla and Sacks 2017; Besley, Burgess, Khan and Xu 2022). Understanding how public administrations function and differ around the world is crucial for identifying the drivers of successful administrations. Recent research in economics, political science and public administration has capitalized on the benefits of direct surveys of public servants to investigate the functioning of the state (for example, Rasul and Rogger 2018; Rasul, Rogger and Williams 2020; Bertrand et al. 2020; Raffler 2022; Harris et al., Forthcoming). This paper presents an initiative – the Global Survey of Public Servants - to collect and harmonize large-scale, comparable survey data on public servants. Questions around how to strengthen the functioning of state administrations pose an immediate measurement challenge: how can differences between public administrations be effectively measured? A valuable response to this has been to generate quantitative indicators based on surveying experts, households and firms, and constructing country-level governance indicators based on their responses (Transparency International 2021; World Bank 2022). These indicators have made important contributions to our understanding of comparative public administration by quantifying perceptions of experts and of those interacting with government (such as citizens and firms) around the world. However, limitations of these indicators exist (Arndt and Oman 2006). For example, country-level governance indicators suffer from the fact that differences within countries across organizations are often larger than cross-country differences (Meyer-Sahling, Schuster, and Mikkelsen 2018), and country-level scores are often not informative to 2 understand cross-country differences in functionally equivalent organizations (Gingerich 2013). As a complement to existing measures of comparative bureaucracy, direct survey measures of the experiences and perceptions of bureaucrats themselves provide an analogous source of data to household and enterprise surveys, but for the public service. This paper introduces the Global Survey of Public Servants (GSPS), a global initiative to harmonize large-scale comparable micro survey data from public servants around the world (accessible at: https://www.globalsurveyofpublicservants.org/). The GSPS was founded by the authors of this paper – researchers at or affiliated with the World Bank Bureaucracy Lab, the Stanford University Governance Project, University College London, and Nottingham University. The intention of the GSPS is first simply to encourage further surveying of public officials around the world. However, publishing data at a more granular level than has been available to date allows for research on questions of the state that are complementary to what has been undertaken to date. To contextualize the GSPS, this paper introduces a conceptual framework of the survey, which reflects the dual academic-practitioner aims of the initiative. Such a conceptual framework is required given how limited the conceptual consensus is in any literature on the drivers of state capacity and administrative quality. The GSPS’s objective is to improve scholarly understanding of how public administrations and states around the world work and to help governments manage public servants better. As such, we require a theoretical lens that allows us to make choices on what measures to include in our effort. The paper then details the methodological approach of the GSPS and current state of play of the GSPS dataset (accessible at: https://www.globalsurveyofpublicservants.org/indicators/). We provide illustrations of how scholars can use the GSPS data to study core topics in comparative public administration 3 differently. We conclude the empirical section by showing how GSPS indicators correlate at the country-level with one of the most widely used sets of governance indicators, the Worldwide Governance Indicators (Kaufmann, Kraay, and Mastruzzi 2010). Literature Review and a Conceptual Framework of the Global Survey of Public Servants Early surveys of public servants by the World Bank were undertaken around 2001 under the umbrella of the ‘Governance and Anti-Corruption Surveys’ initiative (Recanatini et al., 2010). These surveys, as the name suggests, focused on the determinants of corruption and potential policy impacts. In academia, many US-based public administration academics have used the Federal Government’s staff survey, the Office of Personnel Management Federal Employee Viewpoint Survey (OPM FEVS), or complements to it (see for example the body of work of Sanjay K. Pandey). There are a small number of exceptions of research based on surveys from elsewhere in the world, with Grødeland et al. (2001) an early example, but the overall scale of surveying has been far more limited than that of citizens, households or firms (Rogger 2017). Though some governments have undertaken surveys of their officials, regular and consistent surveying within government is a recent phenomenon in most governments (Khurshid and Schuster, forthcoming). Each of these initiatives has, understandably, taken a rather partial view of the determinants of government functioning. In contrast to surveys of private enterprise, for which the economic theory of the firm presents a clear benchmark model, there exists no consensus on the underlying production function of public administration that would guide measurement selection. This is true both theoretically as well as empirically, where the very fact that there has been so little analysis of granular data on public officials has limited the capacity to create stylized facts of public service. As such, this section lays out a conceptual 4 framework for a survey of public servants that we then use to guide our measurement choices. This requirement is particularly important given the comparative nature of the data we are harmonizing. By making public much of the survey data we – and governments – have collected, we have faced the challenges of integrating surveys that were not always implemented with integration in mind. By including a core module of questions consistent across settings and enumerating it to a similar set of central administrators in our own surveys, the scope of integration is increased. Towards this end, we have developed a conceptual framework to define the topics and questions we include in this core module. The conceptual framework of the Global Survey reflects its dual aims: to better understand how public administrations work, and to provide actionable evidence to governments to improve management of public servants. To this end, the core questionnaire of the Global Survey contains measures which reflect core scholarly interests in comparative public administration on the one hand, and government interests on the other. As further detailed in Mikkelsen, Schuster, and Meyer-Sahling (2021), from a scholarly perspective our core questionnaire follows public administration literature in focusing on assessing ‘ideal types’ of modes of governing - “the actual operating modes and administrative arrangements by which rulers govern” (Roth 1968, 156). We distinguish and measure in the GSPS indicators related to three ‘ideal types’: patrimonialism, Weberian bureaucracy, and (new) public management.1 Weber (1978) coined the term ‘patrimonialism’ for an administrative system in which rulers treat the state as private property and govern it as ‘patrons’ through informal connections with – and loyalty and reciprocity of – public employees (their ‘clients’). Patterns of (neo)patrimonial rule persist in many countries in the global south in particular (cf. Fukuyama 2004), underscoring the conceptual importance of patrimonialism for a global 5 survey of public servants. What then are the key concepts for measurement in a survey of public servants that derive from patrimonial rule? In terms of management practices, politicization and ‘personalization’ (nepotism) are central: personnel decisions – such as recruitment or promotions – are decided on political and personal (family and friends) criteria. In terms of employee attitudes and behaviors, corruption (the abuse of public office for private gain) and clientelism (the exchange of state resources for electoral support) stand out – so much so that patrimonialism (or patronage) is often equated with both (Rothstein and Varraich 2017). Weber’s (1978) bureaucratic ideal-type was designed as an antidote to neopatrimonialism: bureaucrats were to become autonomous from political (or other particularistic) influences through merit examinations, job stability, seniority-based promotions, and sufficient pay to reduce corruption to supplement incomes. This was complemented by hierarchy and detailed rules, in line with a Rechtsstaat ideal (cf. Olsen 2006). This set of administrative characteristics was intended to foster key (Weberian) employee attitudes: an esprit de corps, impartiality, political neutrality, integrity, rule following, and expertise. Our core questionnaire thus includes their measurement. While Weberian bureaucracy was an antidote to patrimonialism, it created problems of its own; in particular, a lack of incentives to perform and innovate (Osborne and Gaebler 1992). New Public Management (NPM) sought to address these (Hood 1991). While the term has competing conceptualizations (Dunleavy et al. 2006), at its core it is a “doctrine that the public sector can be improved [through] business concepts, techniques and values” (Pollitt and Bouckaert 2012). For personnel management, this implied a shift to performance: the setting of targets, the measurement of their achievement, and the incentivization of public servants to achieve those targets, for instance through performance pay, temporary job contracts and autonomy to innovate (Osborne and Gaebler 1992). These management 6 practices were to foment a series of favorable employee attitudes and behaviors – such as work motivation, innovation and performance orientation. Again, our core questionnaire therefore includes their measurement. Our core survey thus includes measures that capture key characteristics of patrimonialism, Weberian bureaucracy and New Public Management. They enable scholars to understand to what extent governments and different institutions inside government blend different elements of these ‘ideal-types'. At the same time, to further ensure that our results are actionable, the GSPS module also contains measures that are core to governmental interests when governments themselves undertake surveys of public servants. To identify these measures, we reviewed the topics covered by six governments that regularly conduct government-wide employee surveys: the U.S., Canada, Australia, the UK, Colombia, and Ireland (see Mikkelsen, Schuster, and Meyer-Sahling 2021, for details of this review). Our review finds that governments measure an overlapping set of management practices and employee attitudes. For instance, in terms of employee attitudes, all governments measure employee engagement and organizational commitment, and all but one measure job satisfaction and turnover intention. In terms of management practices, all governments measure employee perceptions of pay, leadership, and performance management. To ensure we measure indicators of practitioner interest, we thus include in the Global Survey core module those measures which at least five out of six reviewed governments include in their surveys. Comparing measures included in government employee surveys and measures of core scholarly interest in academic models of governance, there is some overlap – for instance around performance management or career commitment of public servants. There is also divergence, however, with government employee surveys giving greater pride of place to concepts that are core to organizational psychology – such as job satisfaction or engagement 7 – but seldom referenced in academic ‘ideal types’ of public sector governance; and academic ‘ideal types’ giving greater place to management practices – such as merit recruitment – which are less consistently covered in government employee surveys. To ensure our core module is of interest to both scholars and practitioners, our conceptual framework (and the GSPS core module) thus cover both (see Figure 1). This leads to a questionnaire with 46 questions and a running time of 10 (online) and 15 (in-person) minutes respectively. Figure 1. Core Concepts Measured in the Global Survey of Public Servants 8 Source: Meyer-Sahling et al. (2021) Methodological Approach of the Global Survey of Public Servants Surveying public servants around the world poses an immediate comparability challenge: how to ensure that survey responses can be compared across countries? We take several duties of care, while also being explicit about the limitations of our data. Our intention is also that over time, the initiative will improve the core comparability of the measures as well as learn more about the validity of comparing measures over time, space, and setting. First, in terms of duties of care, our surveys are conducted with comparable survey population frames. All surveys included in the Global Survey of Public Servants target public servants in a broad range of central government institutions – or, in short, central government bureaucrats. Our data does not contain responses from frontline officials such as teachers or doctors, in part as other efforts – such as the OECD Talis survey (OECD 2018) – focus on them. Second, we invite a census or representative sample of bureaucrats in all or a subset of central government institutions in each country to respond to the survey (rather than, e.g., a convenience sample). In a number of cases where comparable sub-national surveys of public administrators were conducted, we include this data in the initiative’s data release. However, the core rationale for inclusion is that the survey has a central government component. Third, we conducted over 150 cognitive interviews with public servants across countries in which we fielded surveys to develop our measures and safeguard a comparable understanding of measures across countries. 9 Fourth, we combine in our questionnaire established measures in the literature where available (e.g., for job satisfaction) with measures developed by ourselves where needed (e.g., to measure whether respondents were hired based on merit). We drew on the aforementioned cognitive interviews and a measurement validation framework (Mikkelsen et al., forthcoming) to develop these measures, and continue to undertake methodological work for further questionnaire improvement (cf. Rogger and Schuster, forthcoming). Our current dataset brings together responses from over 1,300,000 public servants in over 1,300 government institutions in 23 countries in the Americas, Europe, Africa, and Asia (Table 1). The dataset comprises surveys conducted by the authors in 18 countries, as well as published data from 5 countries that conduct their own surveys (UK, US, Australia, Canada, and Colombia). As a result – and as some of our surveys were conducted by GSPS members before our core module was finalized – the country coverage of different indicators varies. Figures 2 to 11 in this paper thus have varying country coverage (we always show the full set of countries available for an indicator). Our online dashboard on https://www.globalsurveyofpublicservants.org/indicators/ enables scholars to compare countries and to assess within-country variation in each indicator – for instance across institutions or demographic groups (e.g., by gender or age group). The data can also be downloaded (at the country, organizational and demographic group level) with confidence intervals (see https://www.globalsurveyofpublicservants.org/datadownloads/). 10 Table 1. Current Country Coverage of the Global Survey of Public Servants # of # of institution respondent Respons Conducted Country Mode Year s s e rate by National Australia Online 2019 102 104,471 77% government National Canada Online 2019 86 182,306 62.30% government Chile Online 2019 66 26,106 43.70% GSPS National Colombia Online 2019 202 25,082 95.20% government Ethiopia In-person 2016 78 2,195 99.50% GSPS Ghana In-person 2018 51 3,343 95% GSPS Albania Online 2017 17 3,690 47% GSPS Brazil Online 2017 18 3,992 11% GSPS Estonia Online 2017 49 3,555 25% GSPS Kosovo Online 2017 62 2,465 14% GSPS Croatia Online 2017 64 6,711 18% GSPS Slovak Republic Online 2017 100 10,817 33% GSPS Guatemala In-person 2019 18 3,670 96% GSPS Indonesia In-person 2012 15 3,903 83% GSPS Liberia In-person 2016 32 2,790 92% GSPS In-person Lithuania (via video) 2021 11 956 83% GSPS Nigeria In-person 2010 94 5,630 100% GSPS Philippine s In-person 2014 7 2,573 100% GSPS 92% (In- person), In-person + 24% Romania online 2019 81 6,037 online GSPS United National Kingdom Online 2019 106 308,556 67% government Ukraine Online 2018 15 1,802 43% GSPS United National States Online 2019 45 615,395 42.60% government Uruguay Online 2021 20 10,232 29% GSPS Total 1,339 1,336,227 63% Before providing illustrative insights based on the indicators to motivate future scholarly work, several caveats about cross-country comparisons using the data are due. 11 First, survey modes across countries differed, ranging from in-person to online to phone. Our evidence suggests that cross-country comparisons are nonetheless possible. Several of us conducted a field experiment in Romania, in which we randomly assigned respondents to complete our survey online or in-person. We found only small mode effects for national averages (Han et al., forthcoming). Second, response rates across countries differed, from 11% to 95%. Our evidence suggests that cross-country comparisons are nonetheless possible. In a field experiment in Romania, we found that varying response rates to an online survey across organizations did not have substantive effects on means (relative to an in-person survey with consistently high response rates across organizations) (Han et al., forthcoming). Third, while our survey populations were consistently central government bureaucrats, practical constraints in collaborating with different governments on the survey imply that the precise set of institutions included in central government varies across countries, as does the coverage of central government institutions. Differences in country means might thus reflect differences in organizational coverage inside governments, though – as a criterion for inclusion in the dataset – we surveyed a broad range of central government institutions. Fourth, for some of the measures, there are differences in precise wording or scales of survey measures – in particular for measures for which our cross-country coverage extends to governments implementing their own surveys. The dashboard and dataset make these differences explicit, for instance by showing the exact question text (translated into English) when hovering over indicators. Wording differences – where applicable – should be kept in mind when comparing country means. Lastly, despite a large set of cognitive interviews, we cannot rule out that cross- country differences stem from differences in language, social desirability biases or meaning 12 across countries (cf. Jilke, Meuleman, and Van de Walle 2015). Particularly culturally sensitive concepts – such as public service motivation – may suffer from scalar non- invariance (Mikkelsen, Schuster, and Meyer-Sahling 2021). They thus require more care when interpreting cross-country differences than factual questions in our questionnaire (such as whether respondents had a performance evaluation in the last two years). While these are important limitations to keep in mind, they of course equally affect many other approaches to constructing cross-country governance indicators. In an expert survey, for instance, experts cannot be randomly sampled, respond to questions in different languages and might exhibit different social desirability biases, among others. We thus believe that our approach nonetheless adds to the available landscape of cross-country governance indicators, and with more granularity than most existing indicators. Moreover, as the initiative develops and the core module is included in a wider range of surveys, some of the abovementioned concerns will be reduced. Uses of the Global Survey of Public Servants How can the Global Survey of Public Servants enrich the study of comparative public administration? To illustrate their utility, this section sketches out six possible uses of the indicators. First, the indicators can be used to nuance popular global stereotypes – at times reflected in the literature – about bureaucracies. One such stereotype is that bureaucrats around the world lack performance incentives: their work effort is not rewarded with, for instance, better pay or greater promotion prospects (see, classically, Osborne and Gaebler 1992). The data from the Global Survey of Public Servants suggest important nuances to this stereotype. In line with conventional wisdom, only a minority of public servants – 31% on 13 average across countries in our sample – believes that work performance matters for pay rises (Figure 2). Figure 2. Performance Pay, by Country (% of public servants indicating that work performance is important for pay rises) Source: Fukuyama et al. (2022) However, contrary to stereotypes about bureaucracies, a majority – 70% on average in our sample – believes that their work performance has at least some importance for their promotion prospects (Figure 3). Public servants may not be as deprived of performance incentives as stereotypes presume, at least in the countries that we analyze. Figure 3. Performance-Based Promotions, by Country (% of public servants indicating that work performance is important for promotion prospects) 14 Source: Fukuyama et al. (2022) Second, the GSPS can – with the appropriate caveats noted above including those related to variation in exact question wording and scale – help understand sharp differences in employee attitudes and management practices across countries. For instance, as illustrated in Figure 4, the share of respondents indicating that it would be difficult to dismiss them or public servants generally varies between 21% (Slovak Republic) to 70% (Brazil). These results suggest sharp differences in the perceived job stability of bureaucrats across countries. Figure 4. Job Stability by Country (% of public servants agreeing that it would be difficult to dismiss them/public servants) Source: Fukuyama et al. (2022) 15 Third, the GSPS indicators can enable scholars to understand change over time in recruitment practices in governments, by assessing how recruitment practices interact with years of service of public servants.2 To illustrate, Figure 5 plots the share of public servants indicating that they found out about their public sector job through a public advertisement in several countries (Chile, Croatia, Romania, Estonia), by years of service of public servants. It shows that in a range of countries, the share of public servants finding out about their first public sector job through an advertisement (e.g., newspaper, government website) has increased sharply. For instance, in Chile, only 18% of public servants recruited more than 20 years ago learned about their first public sector job through an advertisement, relative to 44% of public servants recruited in the last decade. With jobs being publicly advertised rather than only disseminated through word of mouth, transparency in public sector recruitment has arguably increased in these countries, at least in terms of vacancy dissemination. Our data can thus also help assess changes in civil service management and reforms where applicable. Figure 5. Recruitment through Job Advertisements in Select Countries, by Years of Service of Public Servants (% of public servants indicating they found out about their first public sector job vacancy through an advertisement) Source: Fukuyama et al. (2022) 16 Fourth, the GSPS indicators can shed light on variation across countries in the nature and internal organization of the state, for instance in how centralized or decentralized civil services are. To illustrate, Figures 6 and 7 compare variation across organizations inside the governments of Albania and Croatia in pay competitiveness (measured by whether respondents agree that it would be easy for them to find a private sector job that pays better). In Albania, variation in pay competitiveness with the private sector across organizations is relatively limited, with the share of respondents indicating that it would be easy for them to find a better-paid private sector job ranging from 42% to 64% (Figure 6). Figure 6. (Lack of) Pay Competitiveness in Albania, by Organization (% of public servants indicating it would be easy for them to find a private sector job that pays better) Source: Fukuyama et al. (2022) By contrast, in Croatia, this share varies much more sharply across organizations, between 6% and 92% (Figure 7). It reflects much greater decentralization of pay setting in Croatia than in Albania, with different government organizations in Croatia in some cases counting on separate pay regulations. 17 Figure 7. (Lack of) Pay Competitiveness in Croatia, by Organization (% of public servants indicating it would be easy for them to find a private sector job that pays better) Source: Fukuyama et al. (2022) The granular nature of our data thereby allows further nuance in understanding the (de)centralized nature of the state. For instance, within Albania, pay setting is relatively homogenous across institutions, while, in practice, recruitment practices vary sharply. To illustrate, the share of public servants indicating that they were assessed by a written exam in their initial recruitment to the public sector varies across organizations in Albania between 21% and 73% (Figure 8), reflecting the requirement for civil servants to pass a centrally administered exam but not for employees of institutions regulated by the labor code. In other words, within countries, some civil service management practices are more centralized while others vary sharply across institutions. The GSPS indicators provide granular data across countries to help scholars study and understand these nuances. Figure 8. Merit Recruitment in Albania, by Organization (% of public servants indicating being assessed through a written exam in their public sector recruitment) 18 Source: Fukuyama et al. (2022) The perspective of the GSPS is that cross-country analysis and a deeper dive into the variation exhibited across institutions within a single country are complementary lenses by which researchers and public sector managers can understand a specific public administration. Figure 9 showcases how national averages (darker blue bars) compare for the satisfaction with salary across countries. These differences are partly driven by features of the labor markets in those countries. These national averages can be compared to the distribution of satisfaction rates across organizations (lighter blue bars). As such, benchmarking across countries can be an input to an improved understanding of the variation observed within a single government. Figure 9. Salary Satisfaction in the United Kingdom, by Country and Organization (% of public servants indicating that they are satisfied with their salary) 19 Beyond helping scholars understand heterogeneity inside government across institutions, the GSPS indicators also enable scholars to study how the experience of different demographic groups of public servants varies across countries – for instance in terms of generational differences, rank differences, differences by contract type of employees or gender differences. To illustrate, Figure 10 splits pay satisfaction of public servants by gender within each country. The graph shows that gender differences in pay satisfaction are relatively small in many countries. Yet the figure also shows that in some governments – such as Ethiopia – significantly more female employees are satisfied with their pay, whereas in others – such as the Slovak Republic – significantly more men are satisfied with their pay. Figure 10. Pay Satisfaction, by Country and Gender (% of public servants agreeing they are satisfied with their pay and/or total benefits) 20 Source: Fukuyama et al. (2022) The data thus enable scholars to understand systematically how sociodemographic factors like gender differences vary systematically across countries – from employee attitudes such as motivation or job satisfaction, to perceptions of male and female employees of management practices such as leadership. Observing these differences, scholars may seek explanations for why, in some countries, men or women respond more favorably than in others. To our knowledge, this research endeavor – exploring variation across countries in the experience of different demographic groups of public servants – was previously not feasible for scholars at scale. Finally – and perhaps most straightforwardly – our indicators may also help scholars implementing surveys of public servants in specific countries understand how their case countries are similar or different from others in the world in employee attitudes or management practices. We hope the indicators can thus lead to more empirically grounded case selection and generalizability discussions in country-specific research. Relating the Global Survey of Public Servants to Other Governance Indicators How do existing composite governance indicators relate to the GSPS indicators? To illustrate, consider, first, the relation between GSPS indicators and the widely used Worldwide Governance Indicators (World Bank 2022). We do not intend the comparison to be a `validation’ exercise, but rather an exploration of the linkages between the GSPS data and the existing literature. Figure 11 provides scatterplots showing the relationship at the country-level between three GSPS indicators – work motivation, pay satisfaction and perceived job stability of public servants – and the two indicators from the Worldwide 21 Governance Indicators that arguably most closely relate to public administration: Government Effectiveness and Control of Corruption. Figure 11. Global Survey of Public Servants Country Indicators vs. Worldwide Governance Indicators The graphs show that different GSPS indicators correlate differently with Government Effectiveness and Control of Corruption. Greater work motivation of public servants is positively correlated with both greater government effectiveness (r=0.47) and better control of corruption (r=0.35) (greater work motivation is measured by whether public servants agree that they are willing to go the extra mile or perform tasks that are not really required from them, among others). By contrast, pay satisfaction of public servants is negatively – albeit weakly so – correlated with both governance indicators, whereas perceived job stability is (weakly) positively related with corruption control, but negatively related to government effectiveness. 22 While these are merely suggestive correlations (which might additionally be affected by differences in wording across countries), they do underscore that the Global Survey can uncover differences in nuanced characteristics of public administration across countries and enable scholars to explore their relevance for governance outcomes in a way that composite governance indicators cannot and do not predict (as different employee attitudes and management practices in the GSPS indicators are differentially correlated with composite governance scores3). Conclusions and Outlook How do civil service management practices differ within and across governments? How do core attitudes of public servants – such as their motivation or satisfaction – differ within and across governments? This paper introduces the Global Survey of Public Servants, which enables scholars to draw on micro-data from public servants around the world to understand these key questions in comparative public administration. The GSPS can help scholars explore stereotypes about public administrations, assess changes in civil service recruitment, and understand the internal dynamics of governments. For instance, in terms of heterogeneity across institutions, across indicators of public service management, and across demographic groups. They also enable scholars to compare public administrations around the world (with due limitations about measurement invariance and differences in languages and question wording, among others, kept in mind). Existing composite governance scores are less resource intensive to collect and will continue to have advantages in terms of cross-country and intertemporal coverage. At the same time, where the GSPS is implemented, it arguably brings greater nuance and actionability than composite governance scores by enabling more disaggregated assessments of specific management practices in specific organizations or for specific groups of public 23 servants inside government. We are thus confident that the GSPS contributes to the growing set of micro datasets available to study governments around the world (cf. Baig et al. 2021; Charron, Lapuente, and Bauhr 2021; Fazekas and Czibik 2021), and offers a sea change in micro data availability to study comparative civil service management and organizational behavior in particular. While this paper presented the current state of play of the GSPS, the GSPS co- founders (the co-authors of this piece) will continue to conduct surveys of public servants with governments around the world and work with governments which already implement surveys to align their measures with the GSPS core questionnaire module. As such, we hope that the GSPS dataset – and its utility to scholars – will further grow over time in coverage and comparability. We hope other scholars may join the effort, and we are keen to collaborate with colleagues seeking to conduct government-wide surveys of public servants in countries not currently covered by the GSPS. We also highlight the need for undertaking further methodological validation work on the GSPS indicators and on the measurement of public administration more generally (cf. Rogger and Schuster, forthcoming). Understanding the functioning of public administration is complicated by the lack of objective benchmarks in much government activity. Micro-data on the experience of public servants of their work contributes to a more granular approach to understanding the state despite this, and coherence in that data across organizations and countries facilitates more precise comparative work. Our hope is that the Global Survey of Public Servants is a step towards each of these aims, and towards a stronger study of the public service. 24 Endnotes 1. These are, of course, not the only ideal-types advanced by scholars. Since the 1990s, a number of others – such as network governance or ‘digital era governance’ – have been posited (Dunleavy et al. 2006). None of them have been as ‘dominant’ a model as the three we measure, however (Pollitt and Bouckaert 2012). 2. This is suggestive only, as it is based on responses from public servants who were recruited in the past but remain in public service at the time of the survey. Given the low rate of turnover in many public services, however, this may be more credible an exercise than in a comparable private sector setting. 3. Composite governance scores draw, of course, from underlying indicators. Yet, these are themselves typically aggregated. ‘Government Effectiveness’, for instance, aggregates indicators such as ‘Quality of public administration’ or ‘Institutional effectiveness’ (World Bank 2022). 25 Bibliography Agerberg, Mattias. 2022. Corrupted Estimates? Response Bias in Citizen Surveys on Corruption. Political Behavior 44 (2):653–78. https://doi.org/10.1007/s11109-020- 09630-5. Arndt, Christiane, and Charles Oman. 2006. Uses and Abuses of Governance Indicators. Paris: OECD Development Centre. https://www.oecd-ilibrary.org/development/uses- and-abuses-of-governance-indicators_9789264026865-en. Baig, Faisal Ali, Xu Han, Zahid Hasnain, and Daniel Rogger. 2021. Introducing the Worldwide Bureaucracy Indicators: A New Global Dataset on Public Sector Employment and Compensation. Public Administration Review 81 (3): 564–71. doi:10.1111/puar.13355. Besley, Timothy, Robin Burgess, Adnan Khan and Guo Xu. 2022. Bureaucracy and Development. Annual Review of Economics 14 (1): 397-424. Bertrand, Marianne, Robin Burgess, Arunish Chawla and Guo Xu. 2020. The Glittering Prizes: Career Incentives and Bureaucrat Performance. The Review of Economic Studies 87 (2): 626–655. https://doi.org/10.1093/restud/rdz029 Charron, Nicholas, Victor Lapuente, and Monika Bauhr. Sub-National Quality of Government in EU Member States: Presenting the 2021 European Quality of Government Index and Its Relationship with Covid-19 Indicators. QoG The Quality of Government Institute Working Paper Series 2021:4, University of Gothernburg, Gothenburg, Sweden, April 2021. http://hdl.handle.net/2077/68410. Dunleavy, Patrick, Helen Margetts, Simon Bastow, and Jane Tinkler. 2006. New Public Management Is Dead—Long Live Digital-Era Governance. Journal of Public Administration Research and Theory 16 (3): 467-94. https://doi.org/10.1093/jopart/mui057. 26 Evans, Peter, and James E. Rauch. 1999. Bureaucracy and Growth: A Cross-National Analysis of the Effects of ‘Weberian’ State Structures on Economic Growth. American Sociological Review 64 (5): 748-65. https://doi.org/10.2307/2657374. Fazekas, Mihály, and Ágnes Czibik. 2021. Measuring Regional Quality of Government: The Public Spending Quality Index Based on Government Contracting Data. Regional Studies 55 (8): 1459-72. https://doi.org/10.1080/00343404.2021.1902975. Finan, F., Olken, B. A., & Pande, R. (2015). The personnel economics of the state. Handbook of Economic Field Experiments. Fukuyama, Francis. 2004. State-Building: Governance and World Order in the 21st Century. Ithaca: Cornell University Press. Francis Fukuyamaⓡ,Daniel Roggerⓡ,Zahid Hasnainⓡ, Katherine Berschⓡ, Dinsha Mistreeⓡ, Christian Schusterⓡ, Mikkelsen, Kim Sassⓡ, Kerenssa Kayⓡ, and Jan- Hinrik Meyer-Sahlingⓡ. 2022. Global Survey of Public Servants Indicators. Available at https://www.globalsurveyofpublicservants.org/indicators/. Gingerich, Daniel W. 2013. Governance Indicators and the Level of Analysis Problem: Empirical Findings from South America. British Journal of Political Science 43 (3): 505-40. https://doi.org/10.1017/S0007123412000403. Grødeland, Åse B., Tatyana Y. Koshechkina, William L. Miller and Tatiana Košečkina. 2001. A Culture of Corruption? Coping with Government in Post-communist Europe. Hungary: Central European University Press. Han, Xu, Camille Parker, Daniel Rogger, and Christian Schuster. Forthcoming. Determining Survey Modes and Response Rates: Do Public Servants Respond Differently to Online and In-Person Surveys? In The Government Analytics Handbook: Leveraging 27 Data to Strengthen Public Administration edited by Daniel Rogger and Christian Schuster, Chap. 20. Washington, DC: World Bank. Harris, Adam S., Jan-Hinrik Meyer-Sahling, Kimm Sass Mikkelsen, Christian Schuster, Brigitte Seim and Rachel Sigman. Forthcoming. Varieties of connections, varieties of corruption: Evidence from bureaucrats in five countries. Governance. https://doi.org/10.1111/gove.12714 Hood, Christopher. 1991. A Public Management for All Seasons? Public Administration 69 (1): 3-19. https://doi.org/10.1111/j.1467-9299.1991.tb00779.x. Jilke, Sebastian, Bart Meuleman, and Steven Van de Walle. 2015. We Need to Compare, but How? Measurement Equivalence in Comparative Public Administration. Public Administration Review 75 (1): 36-48. https://doi.org/10.1111/puar.12318. Kaufmann, Daniel, Aart Kraay, and Massimo Mastruzzi. 2010. The Worldwide Governance Indicators: Methodology and Analytical Issues. World Bank Policy Research Working Paper No. 5430, Available at SSRN: https://ssrn.com/abstract=1682130. Khurshid, Ayesha and Christian Schuster. Forthcoming. Surveys of Public Servants: The Global Landscape. In The Government Analytics Handbook: Leveraging Data to Strengthen Public Administration edited by Daniel Rogger and Christian Schuster, Chap. 18. Washington, DC: World Bank. Meyer-Sahling, Jan-Hinrik, Christian Schuster, and Kim Sass Mikkselsen. 2018. Civil Service Management in Developing Countries: What Works? Evidence from a Survey with 23,000 Civil Servants in Africa, Asia, Eastern Europe, and Latin America. Report prepared for the UK-Department for International Development (DFID). Mikkelsen, Kim Sass, Christian Schuster, and Jan-Hinrik Meyer-Sahling. 2021. A Cross- Cultural Basis for Public Service? Public Service Motivation Measurement Invariance in an Original Survey of 23,000 Public Servants in Ten Countries and Four World 28 Regions. International Public Management Journal 24 (6): 739-61. https://doi.org/10.1080/10967494.2020.1809580. Mikkelsen, Kim Sassⓡ, Dinsha Mistreeⓡ, Kerenssa Kayⓡ, Jan-Hinrik Meyer-Sahlingⓡ, Francis Fukuyamaⓡ, Christian Schusterⓡ, Zahid Hasnainⓡ, Katherine Berschⓡ, and Daniel Roggerⓡ. Forthcoming. The Global Survey of Public Servants: Measurement Approach. https://www.globalsurveyofpublicservants.org/about. OECD. 2018. TALIS – The OECD Teaching and Learning International Survey. July 22, 2022. https://www.oecd.org/education/talis/. Olsen, Johan P. 2006. Maybe It Is Time to Rediscover Bureaucracy. Journal of Public Administration Research and Theory 16 (1): 1-24. https://doi.org/10.1093/jopart/mui027. Osborne, David, and Ted Gaebler. 1992. Reinventing Government: How the Entrepreneurial Spirit Is Transforming the Public Sector. New York: Addison-Wesley. Pepinsky, Thomas B., Jan H. Pierskalla and Audrey Sacks. 2017. Bureaucracy and Service Delivery. Annual Review of Political Science 20 (1): 249-268. Pollitt, Christopher, and Geert Bouckaert. 2011. Public Management Reform: A Comparative Analysis — New Public Management, Governance, and the Neo-Weberian State. Oxford: Oxford University Press. Raffler, Pia. 2022. Does Political Oversight of the Bureaucracy Increase Accountability? Field Experimental Evidence from a Dominant Party Regime. American Political Science Review 116 (4): 1443-1459. https:// Rasul, Imran and Daniel Rogger. 2018. Management of bureaucrats and public service delivery: Evidence from the nigerian civil service. The Economic Journal 128 (608): 413-446. https://doi.org/10.1111/ecoj.12418 29 Rasul, Imran, Daniel Rogger and Martin J. Williams. 2020. Management, Organizational Performance, and Task Clarity: Evidence from Ghana’s Civil Service. Journal of Public Administration Research and Theory 31 (2): 259-277. https://doi.org/10.1093/jopart/muaa034 Recanatini, Francesca, Grace Morgan and Mohamed Laghdaf Cheikh Malainine. 2010. Governance and Anti-corrupton Diagnostics: Guidance and Tools for Implementation, Monitoring, and Assessment in the Field. Washington, DC: World Rogger, Daniel. 2017. Who serves the poor? Surveying civil servants in the developing world. Washington, D.C.: World Bank Group. Available at https://openknowledge.worldbank.org/handle/10986/12642 Rogger, Daniel, and Christian Schuster, eds. Forthcoming. The Government Analytics Handbook: Leveraging Data to Strengthen Public Administration. Washington, DC: World Bank. Roth, Guenther. 1968. Personal Rulership, Patrimonialism, and Empire-Building in the New States. World Politics 20 (2): 194-206. https://doi.org/10.2307/2009795. Rothstein, Bo, and Aiysha Varraich. 2017. Making Sense of Corruption. Cambridge: Cambridge University Press. Transparency International. 2021. The ABCs of the CPI: How the Corruption Perceptions Index Is Calculated. Last modified December 20, 2021. https://www.transparency.org/en/news/how-cpi-scores-are-calculated. Weber, Max. 1978. Economy and Society: An Outline of Interpretive Sociology. Berkeley: University of California Press. World Bank. 2022. Worldwide Governance Indicators. Accessed July 22, 2022. https://info.worldbank.org/governance/wgi/. 30