PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Contents 1. Introduction................................................................................................................................0 2. The assessment framework .........................................................................................................1 3. Assessing wage bill dimensions ....................................................................................................3 3.1 Wage bill planning .................................................................................................................3 3.2 Wage bill controls ..................................................................................................................8 3.3 Employment levels and distribution ...................................................................................... 10 3.4 Wage competitiveness ......................................................................................................... 15 3.5 Wage equity ........................................................................................................................ 22 3.6 Wage incentives .................................................................................................................. 25 4. Measuring impact ..................................................................................................................... 28 4.1 Fiscal sustainability .............................................................................................................. 28 4.2 Public sector productivity ..................................................................................................... 31 4.3 Labor allocation between the public and private sectors ........................................................ 36 5. Conclusion ................................................................................................................................ 39 Acknowledgements....................................................................................................................... 41 Annex: Definitions and data sources .............................................................................................. 42 References ................................................................................................................................... 44 Figures Figure 1: Assessment framework for public sector employment and compensation................................2 Figure 2: Wage bill as a share of general government expenditures ........... Error! Bookmark not defined. Figure 3: Volatility in annual recruitments, an example from Brazil ........................................................6 Figure 3: Public sector is a large employer globally ............................................................................. 10 Figure 4: Public sector employment varies considerably cross-nationally ............................................. 10 Figure 5: Most public sector employees work in public administration, education, and health. ............. 11 Figure 6: Proportion of public sector workers with no or primary education......................................... 13 Figure 7: The proportion of public sector employees who are female .................................................. 14 Figure 9: Public sector wage premium (compared to all private paid employees, excluding benefits)..... 17 Figure 10 : Access to benefits in the public and private sectors............................................................ 18 Figure 11: Public sector generally pays a premium for education and health professionals ................... 19 Figure 12: Pay compression ratio: public vs private sectors ................................................................. 20 Figure 13: Wages of hospital doctors relative to global median ........................................................... 21 Figure 14: Relative wages of key service delivery staff can vary significantly ......................................... 22 Figure 15: Pay inequity in the Brazilian public sector ........................................................................... 23 Figure 16: Gender inequity in wages in the public sector..................................................................... 24 Figure 17: Summarizing the evidence on performance-related pay...................................................... 26 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Figure 17: There is no strong correlation between the size of the wage bill and fiscal balances ............. 28 Figure 19: The public sector wage premium is not correlated with measures of institutional quality...... 35 Figure 20: Younger public sector workers in low and lower-middle income countries have a higher wage premium.......................................................................................................................................... 36 Boxes Box 1: Brazil’s wage bill micro-data analytics ...................................................................................... 30 Box 2: Pay and employment and public sector productivity in the European Union .............................. 34 Box 3: Public sector wages and the labor market in Ethiopia ............................................................... 38 Tables Table 1: Approaches to measuring public sector productivity .............................................................. 32 1 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 1. Introduction 1. Effective management of public sector employment and compensation is a vital activity of governments with broad implications for fiscal sustainability, public sector productivity, and the competitiveness of the overall labor market. Governments’ expenditures on employee compensation represent approximately 30 percent of their expenditures and wage bill management therefore has obvious fiscal and expenditure efficiency implications. Public sector wages affect the selection, retention, and motivation of public sector workers which in turn impacts productivity, or the amount and quality of government outputs such as infrastructure, regulations, and public services produced per public sector worker. The public sector is a large employer, accounting for 37 percent of global formal employment, and the wage bill represents livelihoods as well as budgetary costs, and changes in government wages are likely to produce significant effects in the whole labor market and the overall economy , including potentially crowding out private sector employment. In many low- and middle-income countries, especially those experiencing fragility, public sector employment is the core ingredient of the political settlement and wage bill reforms have immediate and often severe implications for political stability, peace and security. 2. The objective of employment and wage policies is to maximize public sector productivity in a fiscally sustainable manner and without distorting the overall labor market. Explicit in this objective are difficult technical and political tradeoffs. Wage bill management has traditionally been approached from a fiscal sustainability perspective, often in the context of an economic crisis, with a primacy towards blunt, short-term fixes like across-the-board wage freezes or cuts, and downsizing through voluntary or involuntary schemes that are deemed necessary for immediate fiscal consolidation. Many of these short- term measures have adverse impacts on long-term growth and welfare, and political viability, and often create distortions and perverse incentives.1 For example, freezing basic wages has often resulted in a mushrooming of less transparent allowances and salary supplements that reduce wage bill transparency, harm pay equity, and hurt productivity. Public sector employment and wage practices, therefore, need to give equal primacy to its effects on recruitment, retention, and motivation of workers and ultimately productivity even during a fiscal crisis. 3. This paper aims to provide a framework for conducting public sector employment and compensation assessments that can help World Bank staff, development practitioners and policymakers, apply a consistent and data-informed approach to developing wage and employment policy reforms. Such a framework is necessary given growing debt distress and the need for expenditure rationalization in many of the World Bank’s client countries , but also in light of the renewed emphasis of the importance of state capacity to address global challenges like pandemics, climate change, building human capital, and reducing inequality, all of which require a strong role for the public sector. Inspired by the success of the Public Expenditure and Financial Accountability (PEFA) framework to diagnose the state of public financial management policies and outcomes, the objective of this paper is to provide a robust methodology anchored in data for conducting public sector employment and compensation assessments and for providing a set of quantitative performance indicators on the wage bill, public sector employment, and wages that can be tracked over time. Unlike PEFA, the goal is not at this stage to develop an assessment score of the various aspects of wage bill management, or a detailed toolkit, but rather to 1 IMF, 2016. Managing Government Compensation and Employment: Institutions, Policies, and Reform Challenges . PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 provide more general evidence-based guidance that can help improve the quality of Public Expenditure Reviews, Human Capital Expenditure and Institutional Reviews, and overall policy advice. 4. One challenge for the diagnostic is where to demarcate the boundary between wage bill management and organizational and human resource management. Most of the impacts of employment and compensation policies are mediated by personnel management practices. For example, whether public sector employees are productive depends on merit-based recruitment and performance management as well as appropriate staff numbers and competitive pay. This assessment framework recognizes these overlaps but for the sake of pragmatism and to keep the scope manageable strictly limits the framework to employment and compensation practices. A diagnostic framework for personnel management is a potential area for future work. Similarly, the paper does not cover the important, but complex, issue of public sector pensions as that too would greatly expand its scope, though it does discuss the implications of compensation policies on pensions, and the impact of both on fiscal sustainability. 5. The paper is organized as follows: The next section presents the assessment framework and then goes through each of the dimensions presenting cross-national stylized facts to highlight important elements that should be assessed, proposes a set of guidance questions for the assessment, and key indicators as the evidence base for the assessment. Section 4 then discusses how to assess and measure the impact of pay and employment on the three outcomes of fiscal sustainability, public sector productivity, and labor allocation between public and private sectors. The final section concludes with some suggestions on framing recommendations. Annex 1 details some of the key definitions and data sources used in this paper. 2. The assessment framework 6. The proposed diagnostic framework consists of two parts (These dimensions cover both the institutional and behavioral aspects of employment and compensation practices. The first two dimensions explore the “macro� institutional aspects of the strength of public financial management and human resource management systems for wage bill expenditure planning and execution; and the next four are more “micro� elements likely to influence the behaviors and motivations of individual public employees. 7. Figure 1). The first part assesses six dimensions of public sector employment and compensation practices: 1. The robustness of the wage bill planning process such that annual planning and budgeting is anchored in a fiscal sustainability framework, is informed by comprehensive and accurate data, and is transparent. 2. The strength of wage bill controls to ensure that the budget is executed as planned and that only legally employed staff on the payroll are paid the wages that are due to them. 3. The employment levels and distribution of the public sector workforce such that there is neither over or under-staffing and that the right staff are employed in the right location and in the right positions. 1 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 4. The wage competitiveness of the public sector so that qualified staff are attracted to public sector employment without creating skills shortages in the private sector. 5. Wage equity so that staff in similar jobs with similar skills and similar performance are paid equivalently. 6. Wage incentives in the form of appropriate performance-based career growth and performance bonuses aimed at maximizing worker productivity. 8. These dimensions cover both the institutional and behavioral aspects of employment and compensation practices. The first two dimensions explore the “macro� institutional aspects of the strength of public financial management and human resource management systems for wage bill expenditure planning and execution; and the next four are more “micro� elements likely to influence the behaviors and motivations of individual public employees. Figure 1 : Assessment framework for public sector employment and compensation 9. The second part of the assessment explores the impact of these practices on: 1. Fiscal sustainability: whether the wage bill is putting a stress on public finances, resulting in either rising fiscal deficits or reductions in other important government expenditures, such as capital expenditures. 2. Public sector productivity: whether employment and compensation practices are capacitating and incentivizing the public sector workforce to effectively deliver core government outputs such as infrastructure, services, and regulations. 3. Labor allocation between the public and private sectors : whether the public sector is crowding out the labor market, resulting in skills shortages in the private sector and impeding private-sector job growth. 2 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 10. Given that employment and compensation diagnostics should be “problem- driven�, the linkages between these practices and outcomes are necessary for identifying priority reforms . While the main outcomes of interest will vary depending on country context and greater emphasis can be given to one outcome over another in an assessment, recognizing the importance of all three is necessary for balancing tradeoffs and informing policies. While some linkages between practices and outcomes are theoretically reasonably clear— for example, weak wage bill controls and presence of large numbers of ghost workers hurts fiscal sustainability and productivity — others, such as whether competitive wages improve service delivery, are open empirical questions. 11. Rigorous evidence needs to underpin the assessment and this report will highlight cross- national stylized facts drawing on comparable data sources . Any global benchmarking needs to be used with caution as definitions of the wage bill, public sector employment, and compensation can vary from country to country. While it is common to do cross-national comparisons of the wage bill (either as a share of GDP or government expenditures) based on government fiscal data, differences in which employees are included or excluded from the payroll, and differences in the elements of pay that are accounted for in employee compensation makes these comparisons inaccurate. Countries also have different approaches on what proportion of services should be publicly delivered with public sector personnel as opposed to publicly financed but outsourced to private sector providers for delivery. Other country characteristics, such as size and whether it is a unitary or federal state , can also impact the wage bill and public employment, particularly for service delivery personnel. The Worldwide Bureaucracy Indicators, used here in cross-national comparisons, while not addressing these differences in the scope of the public sector does minimize the risk of inconsistent definitions and measurement errors as it draws on harmonized household surveys using a common taxonomy which enables cross -country comparisons based on similar sources and common definitions of employment and wages. Annex 1 provides more details on this dataset. 3. Assessing wage bill dimensions 3.1 Wage bill planning 12. The wage bill represents a large and less flexible proportion of government expenditures with significant future liabilities. Globally, and noting the difficulties with cross-country comparisons, the wage bill represents approximately 30 percent of government expenditures (Figure 2), with significant variation around this average. In many low and middle-income countries, the wage bill can take up almost half of all government expenditures, and is an even larger share of expenditures for labor-intensive services like teaching and healthcare (for example, teacher salaries represent more than 80 percent of public education expenditures in developing countries). 2 The wage bill as a share of GDP is larger in higher income countries reflecting the bigger scope of governments as incomes rise; but is higher as a share of expenditures in lower income countries (Figure 3). The wage bill is also “sticky� as it is politically difficult to remove public sector workers or to cut their wages. For example, in higher income countries wage bill increases during economic upturns are not matched by similar reductions during downturns. 3 Evaluations of public sector voluntary retirement schemes and downsizing programs reveal that these are generally 2 UNESCO, 2017. UIS Database on Education 3 IMF, 2016. 3 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 not sustainable. 4 These wage bill numbers underestimate the full fiscal costs of public sector workers given the generous pensions benefits that they enjoy. In Brazil for example, the wage bill is 13 percent of GDP and public sector pensions expenditures are another 4 percent of GDP. 5 Figure 2: Wage bill as a share of general government expenditures Source: IMF Government Compensation and Employment Dataset. “ World� is based on the average for 120 countries for which data is available Figure 3: Variations in the wage bill by country income Source: IMF Government Compensation and Employment Dataset. 13. Given the generally large fiscal impact, wage bill management needs to be anchored in a robust, transparent budgetary planning process informed by sound data. Estimating the size of the wage bill, its 4 World Bank 2008. 5 World Bank, 2017. A Fair Adjustment: Efficiency and Equity of Public Spending in Brazil 4 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 trends over time, and decomposing it into changes in employment and wage growth are clearly necessary to establish the aggregate stylized facts, which requires comprehensive and accurate administrative data systems and transparent budgeting and accounting. These data should ideally cover all personnel employed by the government—both central and sub-national governments—either in a centralized and integrated payroll and human resource management system or through a process of consolidation of the different payrolls. Often, however these systems cover only a subset of employees —notably excluding employees paid from extra-budgetary funds— and are inaccurate as human resource data is not linked to the payroll which weakens incentives for agencies to regularly update personnel records given that there are no costs associated with inaccurate data. Budget classifications may also exclude certain categories of employees, like contract workers, from the payroll or may classify de facto salary payments, like honoraria or per diems, under goods and services. For example, in Indonesia honoraria for attending meetings and workshops which were paid through cash outside of the payroll were estimated to be almost a third of the overall wage bill. 6 14. Wage bill decisions should ideally be part of a medium-term budget framework to ensure fiscal sustainability. An IMF review of practices in 42 countries revealed that half of the countries did not integrate wage bill decisions in their budget planning procedures, and that most low and middle -income countries did not undertake medium-term forecasting of the wage bill as part of their budgetary planning. Many countries do have fiscal rules with wage bill expenditure ceilings, though it’s unclear whether these rules constrain decisions. For example, in Brazil the Fiscal Responsibility Act stipulates ceilings on personnel expenditures as a share of revenue, with states that surpass this limit not having access to federal guarantees for international financing. However, steady wage bill growth over the past decade has meant that over half of the states have exceeded this threshold. 7 Many countries also have automatic indexation of wages with inflation which can result in wage bill growth through wage -price spirals and reduce governments’ discretion in adjusting policies to respond to crises. 8 15. Understanding the institutional mechanisms for setting public sector wages is an essential part of the wage bill planning process . Pay setting regimes refers to the laws, regulations, and norms that govern the level of public sector wages, and the processes by which these are set. At a minimum, a review of the pay setting regime should analyze: • The predominant regime of wage determination, whether it is a unilateral decision by the government or though collective bargaining between government and trade unions. • If unilateral decision by government, then the degree of centralization versus decentralization in pay determination —these range from completely centralized systems where pay for all public sector employees is set in a unified pay scale (as for example in Germany, India, and the Philippines) to completely decentralized systems where individual agencies are free to set their own pay scales within a salary envelope (as for example in Sweden). One important element is the extent to which there are different laws and regulations for different categories of public sector employees. In some countries (Japan and Greece, for example) independent 6 World Bank. 2013. Pay Reform in the Indonesian Civil Service 7 World Bank. 2020. Brazil: Public Sector Wage Bill and Personnel Management 8 Holm-Hadulla, Fédéric, Kishore Kamath, Ana Lamo, Javier J. Pérez, and Ludger Schuknecht, 2010, “Public Wages in the Euro Area: Towards Securing Stability and Competitiveness,� Occasional Paper Series (European Central Bank). 5 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 pay review boards play influential advisory roles, although the final decision still lies with the government. • In federal systems, whether the inter-governmental fiscal transfer system can create perverse incentives for over-staffing by local governments and the challenges to comprehensive planning resulting from the shared personnel management responsibilities between central and sub-national governments. • If the pay regime is characterized by collective bargaining, then the union density, or percentage of public sector workers that are unionized, and whether there is a single comprehensive collective agreement for all public sector employees or several, decentralized negotiations can impact wage bill planning. Decentralized negotiations can be organizational (with separate agreements per agency and ministry), occupational (with separate agreements per job group), geographical (with separate agreements for central, regional and local governments), or a combination of these. In some low income and fragile and conflict affected countries donors create parallel civil services and service delivery arrangements that compromise integrated wage bill planning. While these arrangements are motivated by the understandable need to quickly scale up capacity and help deliver donor-funded programs, they risk compromising nascent planning processes and longer-term sustainability, especially when the payroll for these parallel arrangements does flow through the budget. In Afghanistan for example, thousands of donor-funded—both on and off-budget—contracted staff under the National Technical Assistance program undertake core policy and administrative functions and enjoy significantly higher salaries than civil servants. Such programs also create wage bill growth pressures when these contractors need to be absorbed by the public administration once donor funding ends 16. Hiring decisions should ideally be informed by strategic staffing plans that assess needs but are invariably driven by political considerations . There should be a clear line of sight between an Figure 4: Volatility in annual recruitments, an organization’s medium-term strategy, staffing example from Brazil needs, and available resources yet political cycles in hiring, or ad hoc changes in policy, are very common. For example, studies reveal that even in the European Union member countries, in which personnel decisions are presumably more insulated from harmful political considerations, there are political cycles in hiring with increases in public sector employment around election years. 9 Such a political cycle represents decisions that can occur despite the robustness of fiscal planning and generally data-informed policymaking. Wage bill assessments can analyze employment trends from Source: WB staff calculations based on administrative data 9 See Martins, P. S. (2010): “Cronyism,� IZA Discussion Papers 5349 for evidence on Portugal; and Dahlberg, M. and E. Mork. 2008. “Is There an Election Cycle in Public Employment? Separating Time Effects from Election Year Effects,� CESifo Economic Studies 57(3) for evidence on Sweden and Finland. 6 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 administrative data to highlight these inefficiencies. Volatility in hiring is particularly disruptive to sound fiscal management but is common, as illustrated in Figure 4 in a sub-national government in Brazil. 17. One approach that can increase transparency and accountability around these policy decisions is the public disclosure of wage bill information . Some countries produce annual reports on public sector employment and wages for showing progress on various reforms and for informing future policy developments, and also involve academics and think tanks, such as though institutions like fiscal councils, for fact-checking and lend credibility to their analysis. 10 Guidance questions for the assessment 1. What are the main stylized facts about the wage bill: Its size, evolution, and decomposition into employment and wages? 2. Is wage bill management anchored in comprehensive, accurate, and timely data? a. Does the data cover all government employees paid from the budget and from extra- budgetary funds? b. Is the data derived from integrated payroll and human resource systems that are updated every pay cycle? c. Are budget classifications transparent so that all payroll expenditures are accurately recorded? 3. Is wage bill planning integrated in a broader medium-term fiscal framework? Are forward budget estimates derived from models of future wage bill commitments 4. How does the institutional regime for salary setting impact wage bill planning? Are there particular challenges for sub-national wage bill planning due mismatches in staffing decisions and funding between central and sub-national governments 5. Are staffing increases informed by needs and medium-term strategic planning or primarily by political considerations? 6. Is there any public debate or public disclosure around wage bill policies? Key Indicators Indicator Description Data sources Wage bill as share of GDP, Different measures of the aggregate size of the wage IMF Government expenditures, and revenues bill employment and 10 Examples include the Australian Public Service Remuneration Report (https://www.apsc.gov.au/sites/default/files/18330_apsc_-_remuneration_report-accessible-v3.pdf); and the UK: The Annual Report on Senior Salaries (https://assets.publishing.service.gov.uk/government/uploads/syste m/uploads/attachment_ data/file/740064/Fort ieth_Annual_Report_on_Senior_Salaries_2018.pdf) 7 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 (ideally over the past 10 compensation years) dataset; Boost; administrative data Wage bill as a share of An indicator of the how sensitive the fiscal impact of Administrative data domestic revenues (i.e. the wage bill is to external aid flows, especially in low- excluding foreign grants) income countries Wage bill decomposed by Disaggregates the wage bill by the main organizations Administrative data administrative and industry (ministries and agencies) and the main occupational classification groups (e.g. public administration, education, health, and security) that contribute to the wage bill Annual recruitment patterns Provides an indication of whether annual recruitment Administrative data over time. follows strategic needs or is ad hoc and driven by political considerations 3.2 Wage bill controls 18. A key question for the effectiveness of wage bill controls is whether wage bill ceilings are binding for line ministries and agencies. The sanctity of these ceilings depends on whether they are based on authorized positions or anticipated personnel numbers; whether they are set by economic classification (i.e. there is a salary envelope within the overall budget ceiling) and therefore forbid reallocation between wage expenditures and other expenditures or only by administrative organization and provide agencies flexibility in reallocation; and on the power of the ministry of finance. One common problem is wage bill budgeting based on positions instead of personnel which can lead to cash management problems and perverse incentives as ministries and agencies may deliberately maintain vacancies in order use the savings from vacant positions to fund salary top-ups. Ministries and agencies also often violate ceilings by hiring contract workers who are then later converted into staff, or by ignoring the ceilings altogether. In Gabon for example, employment ceilings set in the budget law each year are not adequately adhered to, resulting in recruitments above the ceiling established for each ministry.11 Wage bill assessments, along the lines of the PEFA dimension on budget credibility, can measure deviations between budgeted and actual wage bill expenditures for the general government or the central government, and budgeted and filled positions as indicators of the effectiveness of wage bill ceilings. 19. Another key issue is effective enforcement and control so that only legally employed public servants receive wages and benefits that they are entitled to. Ghost employees —workers who can get paid from the government payroll without being employed by the government, or public sector workers getting multiple illegal salary payments— are a major problem in many low and middle-income countries. A related problem, particularly for teachers and health workers, is workers having multiple p art-time contracts that add up to more than the allowed full-time working hours. The main mechanism to ensure effective wage bill executions, as spelled out in PEFA indicator PI-23, is integrated payroll and human resource management information systems that have comprehensive databases of authorized staff positions, personnel, and payroll with all elements of compensation that is regularly updated with appropriate safeguards and audit trails. For example, Nigeria’s extensive ghost worker problem was 11 World Bank staff assessment 8 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 revealed when the government implemented a digital ID system for civil servants which enabled it to remove over 60,000 ghost workers, saving $1 billion annually. 12 20. Absent these integrated systems, censuses of public sector employees, or risk-based payroll audits, are also an effective control mechanism. Several Sub-Saharan African countries —Central African Republic, Chad, Democratic Republic of Congo, Ghana, Guinea, and Malawi for example —have recently conducted censuses and audits to remove thousands of ghost workers and retirees from payrolls, and to suspend illegal payments, thereby generating significant fiscal savings. 21. A related problem is delays in payments to legitimate employees . Inefficiencies in the public financial management systems can result delays in budget authorization and payments, and salary arrears are a common problem in many low-income and even middle-income countries. For example, in the past Chad community teachers had payment arrears of 10 months and half of newly hired kindergarten teachers in the Philippines received their salaries late. 13 Electronic payment of salaries, and use of mobile payments, has greatly ameliorated this problem. Guidance questions for the assessment 1. Are wage bill ceilings set by the Ministry of Finance adhered to by line ministries? a. Are ceilings set based on positions or actual personnel? b. Do ministries agencies have discretion in spending the savings from vacant positions? 2. How effective are payroll controls as measured by PEFA indicator PI-23? a. Integration of personnel and payroll records b. Management of payroll changes c. Internal control of payroll d. Use of payroll audits 3. Are there problems with ghost workers? Has a payroll audit been conducted to identify issues? 4. Is there are problem of salary arrears and delays in the payment of salaries for public servants? Key Indicators Indicator Description Data sources Deviations between budgeted This indicator measures the extent to which aggregate Government budget and actual general actual expenditure on the compensation of and outturn data government wage bill government workers reflects the amount originally expenditures in the previous approved, as defined in government budget three fiscal years documentation and fiscal reports. 12 Gelb, A., and J. Clark. 2013. “Identification for Development: The Biometrics Revolution.� Working Paper 315, Center for Global Development, Washington, DC 13 Evans, D., and F. Yuan. 2018. The Working Conditions of Teachers in Low and Middle -Income Countries. World Bank 9 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Filled positions as a This indicator measures the sanctity of wage Administrative data proportion of budgeted expenditure ceilings as high percentage of vacant positions positions can create perverse incentives of using savings from vacant positions to finance compensation Ghost workers as a Measures the effectiveness of payroll controls Public employee percentage of total workers census and payroll audits Wage bill arrears as a Measures delays in salary payments Administrative data percentage of total wage bill expenditures 3.3 Employment levels and distribution 22. Wage bill assessments should analyze whether the public sector is over or under-staffed in aggregate. Important metrics in assessing aggregate staffing levels are public sector employment as a share of total, paid (i.e. those working for wage labor, which excludes self-employed workers), and formal sector paid employment (those who have a formal contract and receive benefits like pensions) . The first metric measures the overall labor market footprint of the public sector, while the latter two are better measures of the public sector’s size in the subset of the labor market, namely wage and formal sector employment, that is more comparable. Globally, the public sector is responsible for 16 percent of total employment, 30 percent of wage employment, and 37 percent of formal sector wage employment (Figure 5). 14 The size of the public sector as a share of total employment inc reases with a country’s level of economic development (Figure 6, top panel) that reflects the increasing role of the state in providing social services as incomes rise. There is however, no discernible relationship between country income levels and public sector employment as a share of salaried employment, which suggests that the public sector grows along with private formal sector wage employment (Figure 6, bottom panel). There is considerable cross- country variation around these global averages with public sector shares of total employment ranging from less than 2 percent to over 40 percent, and of paid employment from 10 percent to 70 percent, with four to five-fold variations in these shares at any given income level. Figure 5: Public sector is a large employer globally 14 Total employed individuals are defined as those workers, aged 15 and above, who in the household surveys responded that they had a job in prior week; wage employees are those whose basic remuneration is not directly dependent on the revenue of the unit they work for and are instead paid in wages and salaries, piece work, or in - kind, and therefore, exclude self-employed workers.; and formal sector wage employees are those who also have an employment contract, have health insurance, belong to a union or who are inscribed in a pension program. 10 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Source: Worldwide Bureaucracy Indicators (WWBI) Figure 6: Public sector employment varies considerably cross-nationally 11 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Source: Worldwide Bureaucracy Indicators (WWBI) 23. Equally important is to determine the occupational distribution of public sector employment. Countries have unique legal and occupational classifications of public sector employees that make cross- national comparisons difficult. In many countries all employees are classified as “civil servants�, meaning they enjoy distinct legal protections; in others only management and policy staff are civil servants and others, particularly service delivery staff, have fewer privileges and are governed by the labor code similar to formal private sector employees. The public sector job classification systems also differ across countries. Labor force surveys that use internationally accepted standard industry and occupational classifications can be a basis for cross-national comparisons. These reveal that the public administration (including police), education and health industries employ most public sector employees ( Figure 7) . Figure 7: Most public sector employees work in public administration, education, and health. Source: Worldwide Bureaucracy Indicators (WWBI) 24. Cross- national comparisons of the “right� numbers of education and health professionals are possible given that there are norms, albeit approximate, for benchmarking. Such estimates are more difficult for public administration as it is a broad category that contains a variety of occupations. In education, based on decades of research, a ratio of one teacher for every ten children is considered 12 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 adequate for pre-school and early childhood programs. 15 Student-teacher ratios can be higher in primary and secondary schools; in the OECD countries, for example, the average primary student-teacher ratio is 16.16 The World Health Organization has standards for the number of doctors and nurses necessary to serve a local population. Cross-national comparisons should be complemented with within-country analysis, with appropriate nuances for differences in localities (rural versus urban, remote versus well connected) and the populations to be served, given that the distribution of these public sector workers in public administration, health, education, and security across localities can be unequal. One study found that a quarter of the variation in teacher allocation to schools could not be explained by the number of students.17 While no such standards exist for policy or administrative staff, inter-regional variations can still point to staffing inefficiencies. 25. The assessment should also explore the skill distribution of public sector employees . While on average public sector workers are more educated than their private sector counterparts, there is variation in the proportion of public sector workers with tertiary education across countries. A high proportion of low skilled workers points to the public sector serving a social welfare function and points to potential fiscal savings without compromising public sector productivity through outsourcing of some elementary functions. As Figure 8 shows, the proportion of public sector workers with no or only primary education declines with country incomes, and many low-income countries have on average 20 percent of such low skilled employees, with the proportions rising to as high as 40 percent in some cases. A corollary to this high proportion of low-skilled workers is a high proportion of clerical or support jobs. A functional review of Serbia’s executive branch found that only 70 percent of positions are core functions, while as much as one third are internal administrative support, such as IT, HR, legal, estates, communications, procurement, knowledge management and finance. 18 This poses a much higher burden on the public sector wage bill than in OECD countries, where the trend is to centralize common functions and consolidate corporate services into a Shared Services Centers. Figure 8: Proportion of public sector workers with no or primary education 15 Bowne, Jocelyn Bonnes, Magnuson, Katherine A., Schindler, Holly S., Duncan, Greg J., Yoshikawa, Hiro kazu. 2017. A Meta-Analysis of Class Sizes and Ratios in Early Childhood Education Programs: Are Thresholds of Quality Associated with Greater Impacts on Cognitive, Achievement, and Socioemotional Outcomes? Educational Evaluation and Policy Analysis , Vol. 39, No. 3, pp. 407 – 428. 16 OECD (2014). Education at a Glance: OECD Economic Indicators . 17 Majgaard K, Mingat A. 2012. Education in Sub-Saharan Africa: A Comparative Analysis . Washington, DC: World Bank. 18 World Bank (2016), Serbia: Horizontal Functional Review of Central Government 13 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Source: Worldwide Bureaucracy Indicators (WWBI) 26. The age, grade, and seniority profile of public sector workers can point to skills gaps. For example, prolonged periods of a hiring freezes or disruptions to recruitment because of conflict, can result in missing cadres, as was the case in Cameroon and Sierra Leone. Another problem is a large proportion of older workers, as in Democratic Republic of Congo (DRC), where the inability to finance pensions has meant that many retirees stay on the payroll. An aging public sector workforce is also a problem in high income countries; in Romania, for example, 30 percent of public employees are approaching retirement in the next 10 years, which can have implications for both staff motivation and productivity, and fiscal sustainability given growth in pensions expenditures.19 27. Improving gender equality in public sector employment should be an important policy objective of governments. The public sector’s large labor market footprint means that it can be a strategic lead er in changing norms and behaviors and promote greater gender equality in employment in the overall labor market. Globally 46 percent of all public sector workers are women, and in most countries the proportion of public sector workers who are female is higher than the proportion of private sector workers who are female. Women, however, are significantly under-represented in the public sector in many low and middle-income countries ( Figure 9). There is also considerable horizontal and vertical occupational segregation with women concentrated in certain industries and positions. Globally 74 and 69 percent of the public sector education and healthcare workforce is female respectively, and women occupy only 30 percent of senior official positions. The reasons for this inequality in employment are under studied but, drawing on academic studies on the private sector, likely entail differential caring responsibilities that limit women’s career progression; social norms and attitudes about what type of work women are more suited to; and biases in task assignments so that women are less likely to receive more visible and career- enhancing responsibilities. 20 Figure 9: The proportion of public sector employees who are female 19Under the HRM RAS. 20Czibor, E. 2021. “Gender Equality in the Public Sector: Bridging the Gap Between Experimental Evidence and Practice� 14 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Source: Worldwide Bureaucracy Indicators (WWBI) 28. In many countries there is a dual labor regime for public sector employees with a large contingent of contract workers in service delivery sectors performing similar jobs as permanent employees. Contract teachers and community health workers are very common in developing countries and are different from their civil servant counterparts in that they are paid less, often significantly so, enter employment through less rigorous recruitment criteria, and have time-bound appointments. While some rigorous impact evaluations have shown that in the short-term contract teachers can outperform regular teachers as they are ostensibly more accountable and have an incentive to work harder in order to get civil servant status, there are question marks about the sustainability of this policy.21 Most contract staff, with the backing of employee unions, successfully transition to permanent status and there are significant risks that the lower screening criteria risk more patronage-based hiring. In Indonesia for example, hiring of contract teachers by sub-national governments is highly politically motivated, spiking in election years and increasing in numbers after the introduction of direct elections of district heads. 22 Guidance questions for the assessment 1. Is the public sector over or under-staffed in aggregate and what have been the employment trends over time? 2. Is the public sector over or under-staffed for particular skills and occupations, both in aggregate and in specific organizational and territorial jurisdictions? 3. Are there a large proportion of public sector workers with low educational qualifications or in low-skilled jobs? 4. Is the public sector a gender-equal employer? a. What are the proportion of women in aggregate? b. Is there horizontal and vertical segregation in employment? 21 Bruns, B., D. Filmer and H. A. Patrinos (2011), Making Schools Work: New Evidence on Accountability Reforms, Washington DC, World Bank. 22 Pierskala and Sacks. 2017 15 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 5. Is there a dual labor regime of a large segment of contract workers and what are its implications for public sector employment? Key indicators Indicator Description Data sources Public sector employment as Public sector employment (as defined by International WWBI; labor forces a share of total employment Classification of Status in Employment) as a share of all surveys currently employed individuals Public sector employment as Public sector employment (as defined by ICSE) as a WWBI; labor forces a share of paid employment share of all individuals in paid employment (and surveys excluding non-paid and self-employment and employers Public sector employment as Public sector employment (as defined by ICSE) as a WWBI; labor forces a share of formal paid share of all individuals with formal contracts, social surveys employment security, health insurance or union membership Industry and occupational Decomposition of public sector employment by WWBI; labor forces distribution of public sector industry classification or occupational group surveys employment Student-teacher ratios; per- Provides a measure of the adequacy of key service Administrative data capital health workers delivery staff compared to international norms Educational profile of public Decomposition of public sector workers by WWBI; labor forces sector workers educational qualifications surveys Females as a share of public Female employment within the public sector as a WWBI; labor forces sector employees share of public sector paid employment surveys Females as a share of public Distribution of female employment in the public WWBI; labor forces sector employees by industry sector disaggregated by industry classification surveys Females as a share of public Distribution of female employment in the public WWBI; labor forces sector employees by sector disaggregated by occupational group surveys occupation 3.4 Wage competitiveness 29. Public sector wage levels are clearly crucial for attracting and retaining competent workers, for expenditure efficiency and fiscal sustainability, and for equilibrium outcomes in the entire labor market. Answering the question “does the public sector pay too much or too little� naturally requires an appropriate comparator. There are three potential comparisons that can be done. First, how much does the public sector pay relative to the country’s private sector given that the likely alternative work opportunities are in the domestic private sector.23 Estimating public-private wage differentials is the most natural comparison which has been explored in a very large academic and policy literature. Second, how much does the public sector in that country pay for certain occupations compared to the public sectors of other similar countries? This comparison is useful for public sector occupations that may not have a clear private sector comparator (such as the police) or where there is a natural regional labor market for public sector jobs (such as the European Union) or significant migration (for example, of doctors and nurses). 23In countries with high levels of emigration, like the Philippines and Croatia, the appropriate comparator for certain occupations, like nurses and doctors, will need to include the destination country labor markets. 16 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 And finally, what is the distribution of public sector pay —that is, is there too much or too little pay for certain jobs relative to others within the public sector— which will be considered in the next section under wage inequity. 30. Theoretically, public sector wages should be set at a level relative to private sector wages so that there is a small public sector wage penalty and this penalty should be stable over time.24 Under the theory of “compensating wage differentials�, a particular job should pay more or less than another if there are other non-wage undesirable or desirable characteristics of that job that need to be compensated for. Given the pecuniary and non-pecuniary benefits of the public sector, the most important of which is job security, monetary compensation should be lower than in the private sector (for equivalent workers in equivalent jobs) so that the total compensation accounting for all attributes of jobs is roughly equal in the two sectors. Under this optimal compensation policy, the public sector will be competitive without being distortionary, and there will not be any shortage of skills in either sector. The same principle implies that the wage premium should be annually monitored to ensure that no gap emerges be tween the public and private sectors that can cause a departure from this theoretical optimum. 31. Estimating public sector wage competitiveness compared to the private sector is methodologically complicated. The standard approach in the academic literature is to measure differences in total compensation (basic wages, allowances, bonuses, and monetized in-kind payments like housing and transportation) between the public and private sector for statistically similar workers in similar jobs. Given the demographic differences of workers between the two sectors, this approach ideally requires controlling for observable worker characteristics, such as age, education, work experience, and gender that impact human capital and therefore earnings; accounting for unobserved characteristics such as ability, risk aversion, and public service motivation; and controlling for occupations given that the similar workers can have very different responsibilities in different occupations. A simple raw comparison of average wages in the private and in the public sector is misleading as public sector workers are older and more educated than their private sector counterparts, have different career objectives and motivations, and work on occupations like administration, healthcare, security, and teaching that are not well represented in the private sector with its much higher proportion of agricultural and manufacturing jobs. Given that many public sector jobs are unique and may not have an obvious private sector comparator complicates this analysis. Another challenge is that the elements of total compensation, such as the proportion paid in wages versus bonuses or benefits, can be quite different between the two sectors and may not be accurately measured in labor force surveys. 32. Cross-national wage regressions for labor force surveys reveal that public sector workers on average earn higher wages than observably similar private sector workers. Using a wage regression where wages are a function of observable worker characteristics, such as education, age (a proxy for work experience), gender, location, and the sector of employment (public sector or private sector), public sector workers have approximately 19 percent higher basic wages (excluding allowances and bonus payments) across the 111 countries for which the World Bank has data, with 80 countries having a positive premium, and there is no clear pattern in the size of the premium with country incomes ( Figure 10). This finding holds for gross wages that include employer social insurance contribution and allowances for the 24 Gomes (2016). https://www.youtube.com/watch?v=Uar2IvyLxLI 17 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 27 European Union member states for which data is available.25 Several academic studies, either single- country or cross-country, similarly find public sector wage premiums. 26 Public sector premia in general are likely to be higher globally when allowances and benefits are accounted for as a much higher proportion of public sector workers receive health insurance or pensions, and higher still when non- pecuniary benefits, like job security, are also added (Figure 11). For example, in Indonesia, Pakistan, and Thailand the inclusion of expected pensions benefits (monetized annually) significantly increased the public sector wage premia (for example, from 15 percent to 42 percent in the case of Thailand). 27 Figure 10 : Public sector wage premium (compared to all private paid employees, excluding benefits) Source: World Bank Worldwide Bureaucracy Indicators, latest available data. “World� is based on the average for 111 countries for which data is available 33. The size of the wage premium is sensitive to the choice of the private sector comparator. What matters for workers are their alternative employment prospects when applying for public sector jobs or when considering leaving public sector employment. The relatively few studies that have analyzed movements of workers between the two sectors suggest that the alternative to public sector employment is more likely to be the formal sector for relatively skilled workers, and the informal sector for relatively unskilled workers.28 The average basic wage premium globally decreases to 7 percent if public sector workers are compared only to formal sector workers, with 52 out of 84 countries having a positive premium. Similarly, for specialized occupations, the comparison should not be to the overall private sector but for those specific occupations in the private sector. For example, there is a wage premium for public 25 WB staff calculations based on Eurostat EU SILC data 26 Finan, Frederico, Benjamin A Olken and Rohini Pande. 2017. The personnel economics of the developing state. In Handbook of Economic Field Experiments. Vol. 2 Elsevier pp. 467{514. 27 Whitehouse, E. 2006 . “Reforming Civil Service Pensions in Pakistan �; Palacios R., and Himanshi Jain H. 2020. “Pension Provision in Thailand Rethinking old age income support in a rapidly aging society � 28 Yassin, S., and F. Langot (2018): “Informality, public employment and employment protection in developing countries,� Journal of Comparative Economics, 46(1), 326 – 348. 18 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 sector education and health professionals compared to their private sector counterparts though interestingly the premium decreases with country income levels (Figure 12). Figure 11 : Access to benefits in the public and private sectors Source: World Bank Worldwide Bureaucracy Indicators, latest available data. Based on the average for 128 countries for which data is available 34. How the wage premium varies across different categories of public sector workers can reveal potential inefficiencies in the wage bill . A common finding in several academic studies, and confirmed in the WWBI data, is that the premium is higher for lower skilled workers and for lower wage occupations. Public sector workers with tertiary education have a wage premium of 2 percent as compared to a 9 percent public sector wage premium for workers with primary education. Clerical occupations enjoy higher premia in the public sector than senior managerial occupations. In many countries, the public sector pays a wage penalty for the most skilled employees. 29 Studies have also noted that premiums are higher for entry-level civil servants than senior civil servants, which may be particularly disruptive to the overall labor market. For example, in Brazil the public sector wage premium is 45 percent for new entrants in the federal civil service, and in Indonesia entry level civil servants earn approximately 28 percent more than similarly situated peers in the private sector. 30 Given these high premia, it is not surprising that there is massive demand for public sector jobs, with hundreds of applicants for each vacancy. 29Hospido and Moral-Benito (2016); Hausman, Nedelkoska, and Noor. 2020 30Kuipers, N. 2020. “The Indonesian Public Sector Wage Premium�; Araujo, Felipe A. 2020. “Selection Bias and the Returns to Public-Sector Employment." 19 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Figure 12: Public sector generally pays a premium for education and health professionals Source: WWBI 35. Even if the public sector pays higher wages on average there may be rigidities in pay policies that prevent it from paying competitive wages for high demand occupations . As has been noted in many studies, pay dispersion in the private sector has increased greatly over the past four decades and some occupations like finance and information technology can pay multiples for similarly educated workers than other occupations. Once common metric for wage dispersion is the wage compression ratio which is the ratio of the 90th percentile wage to the 10th percentile wage in the salary distribution. This ratio is lower in the public sector for 70 out of 99 countries for which there is data in the WWBI (Figure 13). Studies show that when the wage structure in the public sector becomes relatively more compressed, it is more difficult to attract and maintain high-skilled workers or skills that are in high demand in the labor market31. 36. These changes in the structure of the labor market prompted moves towards greater pay flexibility in the public sector in the OECD countries, including greater discretion to ministries and agencies in setting pay scales and more individualization of pay. For example, in Chile finance and regulatory agencies have distinct and higher pay scales than the rest of the public administration ; almost 60 percent of staff have individual contracts determined by the agency head in which pay levels can be set at any point of the pay scale; and most ministries and agencies receive a “critical functions� allowance of up to 100 percent of their salary envelope that they have discretion to distribute among a few staff. While such flexibility can better enable agencies to attract and retain unique skills it also creates risks of pay fragmentation and inequity, and underlines the fundamental tradeoff between equity and pay competitiveness at the top of the salary distribution that governments need to manage.32 31 George J. Borjas, The Wage Structure and the Sorting of Workers into the Public Sector , Working Paper 9313, National Bureau of Economic Research, http://www.nber.org/papers/w9313 32 World Bank 2014. Pay Flexibility and Government Performance 20 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Figure 13 : Pay compression ratio: public vs private sectors Note: The 45-degree line represents equal values for the two axes Source: WWBI 37. For certain occupations there may be no obvious private sector and instead the appropriate benchmark may be the public sectors of other countries. Police and security occupations have limited private sector alternatives. Similarly, globally over 77 percent of all workers in the education sector work in the public sector, with higher ratios in low- and middle-income countries. Doctors and nurses have a high incidence of migration and retaining these workers requires tracking the wages for these occupations in destination countries. 33 The International Comparison Program’s wage data enables cross -national wage comparisons for specific occupations adjusted for purchasing power parity (PPP). They reveal for example that South Africa has pays the highest wages of hospital doctors (relative to the global median) among the countries in Sub-Saharan Africa (Figure 14). Such international public sector wage benchmarking for high demand jobs with relevant countries can be a valuable complement to public- private wage comparisons. 33 Antwi, J. and G. Phillips. 2013. “Wages and health worker retention: Evidence from public sector wage reforms in Ghana�. Journal of Development Economics . https://www.sciencedirect.com/science/article/pii/S0304387812000880 21 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Figure 14: Wages of hospital doctors relative to global median Source: WWBI based on ICP data Guidance questions for the assessment 1. Does the public sector over or under-pay compared to the private sector in aggregate and for different industries, occupations, locations, and demographic groups? 2. How does pay for occupations with limited domestic private sector comparators compare to that in the public sectors of other relevant countries? 3. Is there sufficient flexibility in wage practices to better align public sector wages to the needs of specific labor markets to attract and specialized and high demand talent? 4. Does the government use data to model fiscal impact and identify salary mismatches and to inform wage policy? Key indicators Indicator Description Data sources Public vs. private sector wage Most widely used indicator of overall wage WWBI; labor forces premia (controlling for worker competitiveness surveys and job characteristics) Public-private wage premia by Measures the difference between public and private WWBI; labor forces occupation sector pay for specific occupations surveys Public-private wage premia by Measures the difference between public and private WWBI; labor forces education level sector pay for different skill levels surveys Pay compression ratios in the Measures pay dispersion and is an indicator for public- WWBI; labor forces public and private sectors private wage differentials at different points in pay surveys distribution Pay comparisons for Measures pay of a particular public sector occupation ICP data; occupations across public compared to global, regional, or benchmark country administrative data sectors pay for that occupation 22 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 3.5 Wage equity 38. Wage equity— whether workers in similar jobs, with similar skills, similar years of service, and similar performance are paid equally, and differences in salary between different occupations are deemed legitimate —has strong psychological roots and impacts worker motivation and productivity. Studies have shown that wage equity is ingrained in peoples’ sense of fairness and self-worth, that workers compare their wages to their peers in an organization, just as they do to the private sector, and wage differentials that are not perceived to be justifiable can be demotivating. 34 Surveys of public employees conducted by the World Bank reveal that this pay inequity is a major source of work dissatisfaction—for example, 65 percent of civil servants surveyed in the Liberia forestry agency stated that unequal and unfair pay was a source of conflict between staff.35 Unjustifiably high wages for certain jobs can also have significant implications for pension liabilities as many defined benefit schemes link pension payments to highest earned salaries. 39. There are many sources of wage inequity, a common one being a fragmented wage setting regime with different salary laws for different occupational groups, territorial jurisdictions, and even ministries and agencies. This pay differentiation may be a deliberate policy choice, as in several Western European and OECD countries where responsibilities for pay setting have been devolved to ministries and agencies to better enable managers to individualize pay to the particular competencies of staff and to attract high-demand skills to the public sector. Or it may be an accumulation of ad hoc sector-specific changes reflecting bargaining by different interest groups. In Brazil for example, there are hundreds of different pay scales, and associated employee unions, for different occupational groups in the federal government alone. Such fragmentation can cause considerable pay inequity between similar workers and can also create pressure for wage increases as it encourages competitive bargaining between the unions to increase salaries for their respective occupational groups. 40. Wage inequity can also arise with unified pay scales if these are not grounded in a robust job evaluation methodology. Ideally different jobs should be “weighted� based on standard criteria like complexity of tasks entailed and other relevant traits, and positioned relative to each other either on a single pay scale or on separate pay scales for distinct occupational groups ( for example, public administration, education, health, and security). Often job evaluations are not done or were done in the past and are obsolete as public sector employment grew and new positions were created , and therefore relative pay between jobs can become ad hoc and not grounded in objective and credible criteria. For example, university teachers are much better paid relative to primary and secondary school teachers in low and lower-middle income countries, earning almost 4.5 times the wages of government clerks compared to 1.5 times for primary teachers (Figure 15). These disparities are much lower in high income countries. Similarly, nurses are relatively underpaid compared to doctors in low-income countries. Figure 15: Relative wages of key service delivery staff can vary significantly Wages of teachers relative to clerks Wages of medical workers relative to clerks 34 Rosenfeld, J. 2021. You’re Paid what you are Worth: And Other Myths of the Modern Economy ; Card et al 2012; Breza et al 2017. 35 World Bank. 2020. Liberia Forestry Development Authority: An Institutional Capacity Assessment 23 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Source: WWBI based on ICP data 41. A high proportion of allowances can cause pay dispersion even between similar workers in similar jobs. Brazil is a case in point. For example, in the Federal Social Security, Health, and Labor occupational group, gross pay can vary ten-fold for workers with similar levels of experience, which is largely a result of non-performance related payments and not basic pay (Figure 16).36 Many countries also have a high number of allowances or salary supplements, and often these are at the discretion of a minister or agency head that are often abused. For example, it is common to pay public sector workers honoraria for attending board meetings and workshops, often paid in cash and charged to “goods and services� instead of the payroll, which can become a significant source of rents to those handing out the honoraria and create obvious perverse behavior. In Cameroon for example, these per diems for attending meetings were e quivalent to civil servants’ base wages. 37 Figure 16: Pay inequity in the Brazilian public sector Source: World Bank 2019. Note: Each dot is an employee; the horizontal axis is years of service; the vertical axis is wages 42. Women earn less than men for doing the same work. While the public sector has more gender- equal pay than the private sector, there remains a gender pay gap. Globally, women’s average wages are 88 percent of male wages in the public sector, as compared to 74 percent of male wages in the private 36 World Bank (2019b). 37 World Bank 2018. Cameroon Public Expenditure Review 24 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 sector. There is, predictably, considerable variation across countries on female-mean average wage ratios, though interestingly wage inequity persists in the higher income countries (Figure 17, top panel). A major reason, as discussed, for the gender pay gap is occupational differences with women over-represented in senior positions. This occupational segregation persists even if women are over-represented in the public sector. For example, in the state of Parana in Brazil, while women account for more than 60 percent of tenured employees across government, they occupy only 39 percent of leadership positions. The wage gap also persists even after controlling for occupations and worker characteristics like age and education. For example, women working in the education sector have a wage penalty compared to similarly qualified men, and a similar pattern is found for health and public administration (Figure 17, bottom panel). This gender pay gap for similar workers in similar jobs may be outcome of differential access to allowances and salary supplements; different opportunities for promotion within an occupation; and the impact of differential norms on family care responsibilities. Figure 17: Gender inequity in wages in the public sector Source: WWBI 25 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Guidance questions for the assessment 1. Is there “equal pay for equal work�; i.e. do workers in similar jobs with similar skills get similar wages? 2. Is there gender equality in pay in aggregate and across major occupations? Does government have a policy to address gender inequity? 3. What are the sources of wage inequity? Are pay relativities in salary scales derived from robust job evaluations? Are there discretionary elements of pay, such as cash payments or allowances? Key indicators Indicator Description Data sources Public sector pay compression Ratio of the 10 th percentile of wages, which 90 th to WWBI; labor forces ratio provides an estimate of wage dispersion surveys; administrative data Relative wages between Provides an estimate of pay equity given the different WWBI; administrative different occupations responsibilities and demands of different jobs data Within-occupation public Provides estimates of wage differences controlling for Administrative data sector pay dispersion occupational groups and worker characteristics Female-male wage ratio Measures the raw wage difference between men and WWBI; labor forces (public vs private sector) women surveys Gender wage premium Measures the wage differences between men and WWBI; labor forces women controlling for occupations and observable surveys worker characteristics 3.6 Wage incentives 43. The extent to which wage increases are explicitly linked to performance is important for employee motivation and public sector productivity. There is increasing rigorous evidence that unconditional salary increases have, unsurprisingly, no impact on staff performance. In Indonesia, for example, a doubling of teacher salaries had no effects on improving student learning outcomes.38 The two main channels for wage growth that can potentially be conditioned on performance are pay progression linked to promotions and annual bonuses or pay increments tied to the achievement of yearly goals. Both are dependent on the extent to which competencies and performance inform promotion decisions and performance pay, which in turn depends on how objectively performance can be measured and the effectiveness of the broader human resource management processes. While these personnel management aspects are likely be beyond the scope of employment and compensation assessments, what can be explored are some key stylized facts that can provide suggestive evidence of the productivity and fiscal implications of pay incentives. These include the magnitude of promotion-based and seniority based annual wage increases; the expected wage increases over a representative employee’s career; and the coverage and key design features of performance bonuses. 38De Rhee, J., K. Muralidharan, M. Pradhan, and H. Rogers. 2017. Double for nothing? Experimental Evidence on an Unconditional Teacher Salary Increase in Indonesia. Quarterly Journal of Economics 133, 2: 993-1039 26 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 44. Seniority-based promotions and significant annual within-grade pay increases in many public sector bureaucracies weaken pay incentive effects. Useful indicators of this problem are the percentage of staff undergoing annual performance appraisals and the percentage of staff that receive the highest performance ratings. In Liberia for example, only 30 percent of civil servants surveyed by the World Bank stated that their manager had conducted their performance evaluation in the past year; and in Romania 95 percent of staff self-reported receiving the highest rating in their annual evaluation. 39 These high performance ratings are ubiquitous in bureaucracies, though they do not rule out that promotion decisions may still reflect merit and be based on other sources of information than the annual performance evaluation. Many civil services pay scales have substantial automatic annual pay increases which are fiscally costly, limit governments’ flexibility to respond to fiscal constraints , and have no incentives for performance. In Brazil for example, generous automatic annual pay increases imply that most staff can more than double their real wages after 10 years of service .40 45. An opposite and similarly pervasive problem are flat hierarchies and limited promotion opportunities. In some countries public servants quickly progress to higher grades and then exhaust any opportunities for further promotion. In Romania for example, 70 percent of civil servants had reached the highest technical grades and were ineligible for promotion. 41 Limited promotion is common in teaching and healthcare, in part because of the specialized nature of the job. For example, in Bangladesh less than 10 percent of teachers were promoted to a higher position in their careers and almost no teachers were able to double their salaries over a 30-year career in many African countries. 42 46. Performance bonuses can be an important incentive for jobs that have more easily measurable outputs. Performance bonuses are quite rare in low and middle-income countries and face a variety of design and implementation problems when applied. By contrast, two-thirds of OECD countries have some form of performance pay for their public sector. Performance-related pay in the public sector is a controversial topic given the difficulties in measuring outputs, the multi-dimensional nature of work where measuring some activities can incentivize workers to ignore the unmeasured tasks, and risks of favoritism and pay inequity that can result in the absence of objective performance measures. 43 It is also an area of considerable academic research and the available evidence shows that performance incentives can improve productivity for tasks with standardized delivery processes and relatively easily measurable outputs, such as processing of welfare payments, licensing and registration, tax and customs administration, and, more controversially, education and health. Figure 18 summarizes the evidence from rigorous studies. Performance pay for senior civil servants can also complement measures to improve organizational performance management and results-based budgeting.44 Figure 18: Summarizing the evidence on performance-related pay 39 From World Bank surveys of public sector employees 40 World Bank. 2020. Brazil: Public Sector Wage Bill and Personnel Management 41 World Bank. 2019. Romania: Baseline Review of the National Framework of HRM and its Institutionalization 42 Evans, D., and F. Yuan. 2018. The Working Conditions of Teachers in Low and Middle -Income Countries. World Bank 43 Hasnain et al (2014). The Promise of Performance Pay? Reasons for Caution in Policy Prescriptions for the Core Civil Service. World Bank Research Observer 44 World Bank (2014). Pay Flexibility for Improved Government Performance 27 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Source: Hasnain et al (2014) 47. The design features of performance pay schemes that have direct wage bill and productivity implications irrespective are the proportion of public sector staff eligible for receiving the incentive, the probability of receiving the incentive, and the size of the incentive. Universal performance-pay schemes that do not factor the different attributes of jobs, especially the measurability of outputs, are likely to be costlier with unclear productivity impacts than targeted schemes. If all staff receive the performance bonus, as has been common in many countries, then the incentive disappears, and the bonus becomes just another salary supplement that staff are guaranteed to receive irrespective of performance. In the Brazilian federal government for example, two-thirds of the occupational groups have a performance bonus scheme and over 90 percent of staff in each of these groups received the incentive. In some countries, such as Philippines, there is a mandated distribution of rankings so that only a subset of staff are eligible for the reward. Finally, if the performance incentive is too small then it will likely have no impact but if it is too big then it can create incentives for cheating or tensions between staff given the high stakes involved. 48. Another important type of wage incentive is additional payments to encourage health workers, teachers, and administrators to serve in rural areas or hardship locations, particularly in large and sparsely populated countries. Living conditions are often difficult in rural locations and a small, but growing, experimental literature shows that higher salaries can help address these constraints and improve the recruitment and retention of employees in remote areas.45 The size of the incentive though, needs to be large and financial incentives alone are usually not enough of an incentive and need to be combined with non-financial and career incentives to work. 46 From a wage bill management perspective what is important is that financial incentives be targeted to essential workers that are scarce in particular localities, and to demographic groups, like younger doctors and nurses, who may be more willing to serve in hardship locations and would require lower incentive payments to compensate for opportunity costs. 45 Chelwa, G., M. Pellicer, and M. Maboshe. 2018. Teacher Pay and Educational Outcomes: Evidence from the Rural Hardship Allowance in Zambia. ERSA Working Paper 46 Bhatti, Z. and L. McDonald. 2019. Overcoming shortages of skilled civil servants in remote and hardship areas. Governance Notes, World Bank 28 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Guidance questions for the assessment 1. What wage growth is typical for public sector employees in different occupational groups over the course of their career? How of much of this wage growth is dependent on promotions and how much can occur “within grade� 2. Are promotion decisions largely seniority based and/or automatic? 3. Are there performance bonus schemes and if so what occupation groups and percentage of the staff receive them; and how large is the bonus that staff are expected to receive? 4. Is there targeted, significant additional pay for essential workers to serve in rural areas? Key indicators Indicator Description Data sources Percentage of staff receiving Indicator of the robustness of performance-based Administrative data the highest performance promotions and performance bonuses evaluations The size of annual, within- Measures the importance of seniority vis-à-vis Administrative data grade pay increases performance for wage increases Percentage of staff receiving a An indicator of whether performance bonuses are Administrative data performance bonus regular salary supplements or distinguish between good and poor performers The size of the performance An indicator of the magnitude of the incentive Administrative data bonus as a percentage of basic pay The size of rural or hardship An indicator of the magnitude of the incentive Administrative data allowance and percentage of staff receiving it 4. Measuring impact 4.1 Fiscal sustainability 49. The wage bill can potentially have a major effect on fiscal balances but there are no simple benchmarks of the “right� size of the wage bill . The most used metric for estimating the size of the wage bill, the wage bill as a share of GDP, is not a good indicator of fiscal impact given the cross-national heterogeneity in government functions, scope, and size. While the global wage bill average is approximately 9 percent of GDP, it is incorrect to conclude that countries with wage bills below this number, or below some other average for comparable countries, have more fiscally sustainable wage bills than countries with higher averages. Cross-nationally, there is no correlation between the size of the wage bill, either as a share of GDP or expenditures, and fiscal balances (Figure 19). For example, Denmark has one of the highest wage bills in the world at over 17 percent of GDP but has generally achieved budgetary surpluses. A better measure is the wage bill as a share of expenditures and revenues, but even here there is weak correlation with fiscal deficits. Figure 19: There is no strong correlation between the size of the wage bill and fiscal balances Wage bill (% of GDP) and fiscal deficits Wage bill (% of public expenditures) and fiscal deficits 29 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Source: World Bank Worldwide Bureaucracy Indicators, IMF World Economic Outlook 50. Wage bill dynamics within countries are a better predictor of fiscal unsustainability, though the evidence is limited. There are only a few studies that have explored the relationship between the wage bill, fiscal balances, and non-wage expenditures.47 These find that increases in wage bill tend to worsen fiscal balances, and wage expenditures are pro-cyclical, rising during periods of economic growth but not falling as much during downturns due to structural rigidities. One study of 137 countries found that a 1 percentage point increase in wage-bill-to-GDP was associated with a 0.5 percentage point deterioration in fiscal balances. 48 Another study on the European countries found a similar pattern, with a one percentage point increase in the wage bill as a share of GDP is correlated with an increase in the fiscal deficit of half a percentage point. 49 These studies also find that fiscal deterioration is less in high income countries, which are more likely to finance wage bill increases by increasing revenues and reducing other categories of expenditures, in contrast to low and middle-income countries that tend to finance it by increasing deficits. The risk of fiscal sustainability can be severe in resource-dependent countries. In Iraq, for example, approximately 50 percent of oil revenues are spent on the wage bill which, given the country’s high oil dependency, make it very vulnerable to shocks. 50 51. The sources of wage bill growth vary across countries . An IMF review of 20 case studies revealed that the drivers can be increases in wages (Latvia, Moldova, Romania, and South Africa), expansion in government employment (El Salvador and Portugal), or a combination of the two (Kenya and Tunisia). 51 Sometimes the growth is driven by increases in allowances and non-transparent salary supplements as a back-door way to increase salaries after a period of fiscal consolidation, as was the case in Cameroon. Wage expenditures also have a built-in momentum given that public sector salary scales have seniority- based pay increments that result in a natural rate of wage bill growth even if employment is held steady and there are no across-the-board salary increases. The example of Brazil is indicative, where micro-level data allows for modeling scenarios that can decompose the effects of wage bill increases due to additional hiring from those due to staff moving up the pay scale with increasing years of service. In some Brazilian states, for example, the wage bill increases by over 2 percent in real terms annually even when staffing 47 IMF, 2016. Managing Government Compensation and Employment: Institutions, Policies, and Reform Challenges . Eckhardt, S. and Z. Mills, 2014. What Goes Up Must Come Down. Cyclicality in Public Wage Bill Spending. World Bank Policy Research Working Paper. Dybczak, K. and M. Garcia-Escribano. 2019. Fiscal Implications of Government Wage Bill Spending. IMF Working Paper 48 Dybczak, K. and M. Garcia-Escribano. 2019 49 Eckhardt and Mills, 2014 50 51 IMF, 2016. 30 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 levels are fixed and there is a one-to-one replacement of relatively higher paid retirees with relatively lower paid new hires, and annual nominal increases to the salary scale are limited to inflation. 52 52. The type of wage setting arrangements can have implications for fiscal sustainability. Although the studies are limited, the tentative findings suggest that the risk of “fiscal drift� (an increase in wage bill and an increase in budget deficit) is positively correlated with union density (the percentage of public servants who are members of a union), and the chances of fiscal tightening (reductions in the wage bill and deficits) during an economic recession, as well as during election years, decrease with higher union density. Decentralized collective bargaining agreements where sectoral unions negotiate with their respective counterpart agencies as opposed to centrally negotiated collective bargains with “peak unions� (representing unified smaller unions) likely have higher transaction costs and compromise wage bill planning and the ability of governments to reduce wage bills in the face of crises. 53. Wage bill modelling is invaluable for estimating the risk of fiscal unsustainability and for quantifying the savings from different reform options. Many governments have integrated human resource and payroll systems that they use for undertaking transactions, but which are also a vital, but often under-used, resource for wage bill planning. The micro-data from these systems can form the basis for simulations that estimate the fiscal impact of different scenarios, such as freezing salaries or decreasing the rates at which salaries are readjusted; reducing the size of annual pay progressions or increasing the time interval between them; reducing staff replacement rates, a hiring freeze; and reducing Box 1 : Brazil’s wage bill micro-data analytics The World Bank gained access to vast amounts of individual level administrative data to analyze the Brazilian federal and state wage bills, showing its immediate impact on the country’s fiscal outlook. For the federal government, the analysis revealed that: • Freezing wage for 3 years would yield annual savings equivalent to 25 percent of the primary deficit. • Lowering starting salaries by x% (the federal civil servants receive almost double the wages of private workers with similar characteristics) of would provide savings equal to 13 percent of the primary deficit. • Limiting performance-based payments aimed to civil servants in the top of performance distribution (as opposed to the 90 percent that normally receive them) also yielded significant savings [quantify]. • The increase in the time interval to become eligible for career progression could also be another source of important fiscal savings [quantify] The WB’s analysis received significant media attention that was welcomed given that most of the debate used to revolve around anecdotal evidence as opposed to data. This attention focused on certain elements of the findings, in particular the average retirement age of public servants, the wage premium, and the automatic wage increases in the public sector. The Ministry of Economy is finalizing an administrative reform that is heavily informed by the World Bank report, which would require a constitutional amendment. Source: World Bank. 2020. Brazil: Public Sector Wage Bill and Personnel Management 52 WB staff calculations based on micro-level payroll data 31 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 pay dispersion. Brazil is a good example of where the World Bank team has worked with the government to provide analysis using? wage bill modelling (Box 1). Guidance questions for the assessment 1. What are the wage bill dynamics and how are changes in the wage bill being financed— i.e. through deficits, raising revenues, or reducing other expenditures? 2. What are the main drivers of wage bill growth? 3. What are the wage bill projections for baseline and different policy option scenarios based on wage bill modelling? Key indicators Indicator Description Data sources Correlation between wage bill To measure wage bill dynamics and fiscal sustainability Budget data growth and fiscal balances Correlation between wage bill Provides an indication of whether the wage bill crowds Budget data growth and changes in capital out other spending and non-wage recurrent expenditures Decomposition of wage To understand the sources of wage bill growth Administrative data growth: employment and wages 4.2 Public sector productivity 54. A key question is whether the public sector workforce is performing well and delivering high quality infrastructure, services, and regulations, which is a question of public sector labor productivity. Productivity measures the efficiency with which inputs, like labor, are converted into outputs, and is a more precise and economically meaningful concept than “performance� since presumably performance can be 𝑂𝑢𝑡�𝑢𝑡𝑠 𝑇𝑜𝑡𝑎𝑙 𝐹𝑎�𝑡𝑜𝑟 𝑃𝑟𝑜𝑑𝑢�𝑡𝑖𝑣 𝑖𝑡𝑦 = improved by spending more while productivity measures 𝐴𝑙𝑙 𝐼𝑛�𝑢𝑡𝑠 whether more is produced and delivered for a given 𝑂𝑢𝑡�𝑢𝑡𝑠 𝐿𝑎�𝑜𝑟 𝑃𝑟𝑜𝑑𝑢�𝑡𝑖𝑣𝑖𝑡𝑦 = wage bill. For public employment and compensation 𝐸𝑥�𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒 𝑜𝑛 𝐿𝑎�𝑜𝑟 assessments, the main measure should be labor productivity, though a more commonly used metric in academic studies is total-factor productivity which is the ratio of the currency-value of outputs to the currency-value of all inputs. Productivity is difficult to estimate for the public sector, mainly because of the difficulty in defining and measuring government outputs, particularly for functions such as defense, foreign relations, and environmental protection, and for calculating prices for services for which there are no market transactions. 55. Given these difficulties, a diversity of approaches that go beyond the textbook definitions of outputs necessarily must be used, but which nevertheless provide useful proxies for productivity . These 32 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 alternative measures can be distinguished into `macro’ approaches, which provide information at the level of an organization, sector, or service as a whole; and `micro’ approaches, which can be applied to the individual employee, task, project, and process (Table 1). 53 A common macro approach is to use service delivery outcome indicators per service delivery staff, which, given the high proportion of education and health personnel in the public sector, provides a reasonably comprehensive measure for a large segment of the public sector. These measures of productivity include standardized test scores in the education sector (such as the OECD's Programme for International Student Assessment (PISA) divided by sectoral wage expenditures or the number of teachers for a measure of per-teacher productivity mortality rates or vaccines administered per health worker; and health outputs like number of consultations or patients discharged per doctor or wage spending in the health sector. The use of outcome measures is not without its flaws as it typically includes factors that are beyond the control of the public organization or public official (e.g. quality of support outside of school for test results, and lifestyle choices for mortality rates, family and housing conditions and other demand-side factors). Budget execution rates can be another useful indicator to measure the productivity of public administration employees across a variety of sectors and organizations, with the caveat that fast spending doesn’t necessarily imply good spending. 54 Table 1 : Approaches to measuring public sector productivity Macro (organization, sector, whole public service) Cost-weighted output (Atkinson, 2005) Service delivery indicators Budget execution rates Micro (employee, task, process) Revenue collection Doing Business indicators Procurement outcomes Staff and user satisfaction Subjective assessments (by employees, by stakeholders) Independent observers (`mystery shoppers’) and process productivity The knowledge/capacity of public officials Source: Somani, 2021 56. Micro approaches can be a good complement by providing productivity estimates for large and important organizations like revenue and customs agencies, infrastructure agencies, and social security 53, 45 Somani, R. 2021. Public Sector Productivity: Why is it Important and how can we Measure it? World Bank, Equitable Growth, Finance, and Institutions Insight 33 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 administrations. Examples of outputs include Doing Business scores as a proxy for the productivity of public administration personnel involved in regulations; revenue collection and tax audits conducted per revenue agency staff; project-completion rates for local public-infrastructure projects; organization-level road-construction-completion information; and individual employee-level case completion times in the social-security administration.55 Box 2 provides an example of applying some of these macro and micro approaches to the European Union member states. 55(Rasul and Rogger, 2018); (Rasul et al, 2020); (Lewis-Faupel et al, 2016); (Fenizia, 2020); (Dunleavy and Carrera, 2013). 34 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Box 2 : Pay and employment and public sector productivity in the European Union Below, we present trends in proxies of public-sector productivity for the EU as a whole, based on direct measures of public- sector output. We index the 2010 value at 100 and present the relative changes in the indicator over time. Doing Business: Labor Productivity (per Worker) Doing Business: Total Productivity PISA Reading: Total Productivity 140 140 140 130 130 130 120 120 120 110 110 110 100 100 100 90 90 90 80 80 80 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Life Expectancy: Total Productivity Hospital Discharges: Labor Productivity (per Worker) Revenue Collection: Total Productivity 140 140 140 130 130 130 120 120 120 110 110 110 100 100 100 90 90 90 80 80 80 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2010 2011 2012 2013 2014 2015 2016 2017 2010 2011 2012 2013 2014 2015 2016 2017 The data is sourced from the World Bank’s Doing Business Database, the OECD’s PISA Database, and Eurostat. The top -left figure shows the trend in the Doing Business Score (DBS) divided by the total number of civil servants working in the central administration, as a proxy for labor productivity in the area of business regulation. The top-middle figure shows the DBS divided by government expenditure on general public services, as a proxy for total factor productivity in business regulation. The top-right figure shows PISA reading scores divided by expenditure on education; the bottom-left the life expectancy at birth expenditure on health; the bottom-middle hospital discharges over the number of hospital staff; and the bottom-right the ratio of government revenue to the number of civil servants in central public administration. Bearing the caveats around such measures (presented above) in mind, these trends are still informative of important trends in the quality of public services in the EU, especially when the trends in outputs (numerators) and inputs (denominators) are analyzed separately. The evidence suggests a gain in productivity in business regulation; a decrease in total-factor productivity in social sectors, but evidence of a constant rate of labor productivity over the same period; and evidence of a substantial increase in revenue-collection productivity. Source: WB staff calculations based on Eurostat, Doing Business, and PISA data 57. In the absence of productivity measures, public employment and compensation assessments can explore correlations between the wage bill and expert-based measures of institutional quality. While we may expect better-paid public employees to be more motivated and less corrupt, or to have less of an incentive to supplement their salaries with bribes, cross-national estimates, using the Worldwide Governance Indicators (WGI), reveal no clear relationship between the public sector wage premium and these measures of the quality of governance (Figure 20). In fact, some studies find that nepotism and 35 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 corruption is higher in countries with larges public sector wage premia, suggesting that the higher value of public sector jobs can create opportunities for rent-seeking in recruitment. 56 Figure 20: The public sector wage premium is not correlated with measures of institutional quality Wage premium and Government Effectiveness Wage premium and Corruption Source: World Bank Worldwide Bureaucracy Indicators, World Governance Indicators Guidance questions for the assessment 1. Is total labor expenditure in the sector (including staff wages, outsourced labor costs, and external consultancy fees) correlated with sector-specific improvements in public-sector outputs and outcomes, so that productivity measures are stable or increasing within sector over time? 2. Using subnational data, if available, are there any outliers in the current distribution of total labor expenditure, public-sector output measures, outcome measures, or measures of productivity? Are there any outliers in the trends of these subnational measures (e.g., recent rapid growth or declines in a particular subnational entity)? What might be the underlying features explaining the existence of such outliers? 3. Are total labor expenditures and public sector wages correlated with measures of institutio nal qualityover time? Measures of institutional quality include the Worldwide Governance Indicators, perceptions of the quality of governance through household surveys (e.g., Eurobarometer, Afrobarometer, or World Values Surveys), the OECD’s Government at a Glance Indicators, the Quality of Government Index, or the World Bank’s Doing Business Indicators? With household level data and any other subnational data on measures of institutional quality, are there any outliers across subnational entities in the current distribution of institutional quality or outliers in the trends of measures of institutional quality? Key indicators Indicator Description Data sources 56Rijckeghem, C. V., and B. Weder. 2001. “Bureaucratic corruption and the rate of temptation: do wages in the civil service affect corruption, and by how much?,� Journal of Development Economics, 65(2), 307 – 331. 36 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Correlation between wage-bill To measure trends in public-sector productivity and Budget data (wage-bill growth and public-sector the impacts of wage-bill expenditure on desired growth) outputs and outcomes outputs/outcomes Administrative data (e.g., agency annual reports and statistics agencies for outputs and outcomes) 4.3 Labor allocation between the public and private sectors 58. Public sector employment and compensation policies can have major impacts on the entire labor market, particularly in low and middle-income countries in which the public sector is the dominant formal sector employer. As noted earlier, globally the public sector accounts for 38 percent of formal employment, reaching 48 percent and 49 percent for lower-middle and low-income countries, respectively. This large labor market presence, combined with public sector wage premia and job security, can lead to unemployment, particularly for fresh university graduates who queue for public sector entry- level job openings and reject private sector job offers. There is plenty of anecdotal evidence from even high income countries of thousands of people applying for a few public sector vacancies. 57 Box 3 provides an example of these labor market effects of public sector employment and compensation in Ethiopia and similar conclusions have been drawn based on research on the Colombian and West African labor markets.58 59. Analyzing these labor market effects is technically complicated, but assessments can explore a few stylized facts that can give an indication of these impacts. Given that public sector turnover is low, and vacancies are more likely to occur for entry Figure 21: Younger public sector workers in low and level positions, the public-private wage gap by lower-middle income countries have a higher wage age can provide evidence of the whether public premium sector wages are driving youth unemployment. Cross-nationally we find that the youngest cohort of public sector workers in low and lower- middle income countries — those in the 15 to 24 age group— have a higher wage premium compared to older cohorts which suggests that the distortionary effects of relatively high public sector wages are more pronounced for the youth ( Figure 21 shows the data compared to formal sector private workers). Government administrative data can also be used to analyze these impacts by measuring the number of applications per job opening, with inordinately Source: WWBI high numbers also suggestive of these 57 (Burdett, 2012); (Gomes, 2015, 2017); (Bradley et al, 2017) 58 See Albrecht et al (2017) for Colombia and Girsberger and Meango (2018) for West Africa 37 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 distortions. Complementary data from investment climate assessments or enterprise surveys can be used to assess whether there are skills shortages in the private sector. 38 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Box 3 : Public sector wages and the labor market in Ethiopia The public sector is a highly attractive employment prospect in the Ethiopian labor market and favored towards those with a tertiary level of education. Wages are 50 percent higher in public administration than in services, the second best-paid industry; and individuals working in public administration are 3 times more likely to have a tertiary education than those working in services, the second-highest. As a result of this large wage premium, as opportunities to access tertiary education grew (through an expansion in local public universities), many individuals left the private sector (mainly from service-based jobs) to search for public employment (see top figure). Furthermore, it was only in localities where the public sector was a relatively larger employer (improving the chances of landing a public-sector job) and where the public-sector wage premium was larger (making it more attractive to join the public sector, and hence to undertake a tertiary education in the first place) where individuals chose to enter higher education after a local university was established (bottom figures, left and right, respectively). For example, in localities where the public sector accounted for 50 percent of local employment, tertiary education attainment increased by almost 25 percentage points when a new public university was establ ished. Tertiary education attainment increased much less in districts where the public sector was a smaller employer. Similarly, in districts where the public sector paid double the private sector, local tertiary education attainment increased by 13 percentage points, but much less in areas where the public sector was less lucrative. Together, this provides evidence that the number and wages of public-sector jobs impacts: (i) education decisions (the types of skills and human capital in an economy); and (ii) private-sector labor supply. Estimated Effects on Employment 4% 3.5% 3% 2% 1% Services 0% Public Administration -1% -2% -3% -4% -3.7% -5% Source: Somani, R. Forthcoming. The Returns to Higher Education and Public Employment. World Development; and author’s calculations. 39 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Guidance questions for the assessment 1. Is the public sector crowding out the labor market, impeding private-sector job growth (for specific skills, occupation, demographic, and geographical groups)? 2. Is the public sector having a distortionary impact on younger cohorts, particularly fresh university graduates, causing them to queue for public sector jobs and reject private sector jobs? Key indicators Indicator Description Data sources Public sector wage premium Provides an indicator of relative supply of young WWBI; labor force by age cohorts workers to the public and private sectors surveys Number of applications per Provides an indicator of queuing for public sector jobs Administrative data position (for a sample of positions) Correlation between public- To identify whether public-sector employment or WWBI; labor force sector employment growth or labor demand is crowding out private-sector labor surveys; recruitment patterns and demand or job growth (and the most-affected groups) administrative data private-sector employment growth or recruitment patterns (by skill, demographic, occupation, and geographical groups) Trends in educational profiles To understand trends in the skills and skill demands of WWBI; labor force of public-sector workers the public sector relative to the private sector (the surveys relative to private-sector public-sector education premium) and whether workers (by demographic, specific skill groups are increasingly drawn to or away occupation, and geographical from the public sector groups) Correlation between public- To understand how the public-sector wage premium is WWBI; labor force sector wage premium and impacting labor supply to the public and private surveys relative employment growth sectors in the public sector (by skill, demographic, occupation, and geographical groups) Skills shortages in the private To understand if skills shortages are a constraint for Enterprise surveys and sector (from firm surveys) the private sector Investment Climate Assessments 5. Conclusion 60. This paper has aimed to provide a comprehensive framework for conducting public sector employment and compensation assessments to help develop evidence-based answers to three policy questions that governments care about: how can employment and compensation policies contribute to sound fiscal management, increase public sector productivity, and improve the competitiveness of labor markets and help grow jobs? The importance of these policy questions will vary by country context. Most governments pay more attention to their wage and employment policies in times of fiscal distress and 40 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 therefore, the fiscal sustainability impacts are likely to remain paramount in such assessments. What this paper has emphasized is the importance of a holistic approach since the wage bill is qualitatively different from other government expenditures and other inputs in the government production function, and assessments need to also explore the productivity and labor market impacts. Understanding all these dimensions are necessary for a rigorous diagnostic and for identifying feasible yet meaningful reforms. 61. A big reason by pay and employment reforms are difficult is because of political economy factors, but these have been only touched upon in the framework. With some exceptions, such as on the political cycle in hiring, the assessment framework is largely technical, mainly because political economy factors are country specific and difficult to identify and standardize in a set of diagnostic questions. Political economy factors though are paramount, either explicitly so as in the role of trade unions, or implicitly as public sector employees are in most countries a powerful stakeholder and have a significant voice in what reforms are on and off the table. The country-specific applications of the framework will need to enrich the technical analysis with the underlying political economy factors that result in the status quo practices, and what reforms are politically feasible or require a change in the political equilibrium to be implemented. 62. The data needs are challenging but these assessments can be an impetus for governments to invest in better data and to better use the data that they do collect. Sometime robust data is not available but more often the data that is available is not used. The WWBI provide robust cross -national data, but there are gaps in country coverage due to either the absence of labor force surveys or restrictions in their use for research and policy purposes. Assessments can be used to advocate for improving national statistical capacity or for making the data that is collected publicly available with appropriate safeguards to protect sensitive individual data and abide by inte rnational data privacy standards. Many governments have invested heavily in payroll and human resource management information systems, often with WB and development partner support, but use these systems largely for conducting transactions rather than as a basis for informing policies. These administrative systems provide a rich source of data and can help governments devise the flexible and targeted reforms that are needed and to monitor the impact of these reforms. 41 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Acknowledgements This note was prepared by a team consisting of Zahid Hasnain, Faisal Baig, Ravi Somani, Wouter van Acker, Anca Butnaru, and Daniel Ortega. For the World Bank staff interested in this topic and wishing to explore relevant World Bank cases and relevant literature, please contact govgplearns@worldbank.org 42 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 Annex: Definitions and data sources This report uses the standard definitions of the “general government� and “public sector� in the analysis of the six employment and compensation practices and their impacts. As the figure below depicts, the general government are all institutional units controlled directly, or indirectly, by central and subnational governments and includes the wage bill of all workers employed by these units (e.g. public administrators, service delivery personnel, military and security personnel). The public sector is general government plus public or state-owned corporations. While limiting the analysis to the general government is likely appropriate and sufficient for most diagnostic assessments, in some countries the broader consideration of the public sector may be necessary if there is a large state-owned enterprise sector that has significant claims on the general government budget. Moreover, many household and labor force surveys do not allow for an accurate disaggregation between general government and public enterprises and therefore much of the analysis, particularly public-private comparisons, requires the broader definition of the public sector. Public Sector Public General Corporations Government Central State Local Government Goverments Governments The main data source used for cross-national analysis is the Worldwide Bureaucracy Indicators (WWBI). The objective of the WWBI is to provide comprehensive, cross-national data on public sector employment and compensation to better enable researchers, development practitioners, and policymakers identify and implement evidenced-based reforms. The WWBI country-level indicators are constructed from nationally representative household surveys that are designed and implemented by national statistical agencies. The WWBI are based on microdata from 909 labor force and household welfare surveys, which translates to 53 million unique survey observations and consist of 112,919 estimations across 192 indicators for 202 countries and territories between 2000 and 2018. The WWBI encompass five categories of variables: 1. The size and demographics of the public and private sector workforces (107 indicators) 2. Public sector wage premiums (39 indicators) 3. Relative wages within the public sector (35 indicators) 4. Gender pay gaps (9 indicators) 43 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 5. The public sector wage bill (2 indicators) The demographics of public and private employment track key characteristics including the size of the public sector workforce (in absolute and relative numbers), their age, and distributions across sex, rural and urban locations, academic qualifications, wage quintiles, industry categories and occupational groups. The indicators on public sector wage premiums which capture the overall competitiveness of public sector wages (compared to the private sector) as well as the decomposed public-private wage differentials by sex, academic qualifications, industry category and occupation groups. Indicators on pay compression ratios present the relative wages of the top and bottom earners in the public and private sectors, the ratios of wages for employees of occupational categories in the public and private sector, the relative wages of key occupations within the public sector, and the cross-country comparisons of the compensations of public sector workers by occupations. Indicators on gender pay gap compare the wages of females to their male colleagues in the public and private sectors as well as by industry of employment. Finally, indicators on the relative size of the wage bill offer a glimpse into the structure and affordability of the public sector within the larger economy. The wage data in the WWBI denotes the income associated with the occupation of employment used in the analysis (which the individual dedicated most of their time in the week preceding the survey) and excludes both bonuses, allowances, and other in-cash/-kind payments from the same job as well as all additional sources of income (from other jobs) or investments and transfers. Due to the almost complete lack of information on taxes, the wage from primary job is not net of taxes. For all those with self- employment or owners of own businesses, this corresponds to net revenues (net of all costs excluding taxes) or the amount of salary withdrawn from the business. 44 PUBLIC SECTOR EMPLOYMENT AND COMPENSATION - A FRAMEWORK (6 References 45