Click here for search results


Establishment Control & Pay Determination

Subtopics on this page:  

Public Administration Salary Setting: Principles
and Mechanisms for Satisfying Them

Governments have devised various institutional arrangements to ensure that the right number of public sector staff is hired, at the correct grades and with the right mix of skills for the tasks at hand. This is conventionally referred to as establishment control. Governments have developed public service pay scales that provide a consistent set of rules for matching personal emoluments to staff grades.

This web page describes the traditional model of establishment control, some common critiques and some recent developments, and locates pay within the context of total rewards. The web page concludes with some consideration of sequencing in the reform of establishment control and pay systems.

The traditionalmodel of "establishment control"                            

Under the traditional model, departmental organizational structures are agreed with the central personnel office or equivalent in Anglophone countries and with the Finance Department in Francophone countries (see the page on central agencies). Approval is then given to a hierarchical staff structure and complement for each department, expressed in terms of authorized positions, each with a designated grade level. This constitutes an establishment list, and is the basis for both budget making and personnel recruitment and promotion. The finance ministry prepares the budget following confirmation by the central personnel office that spending proposals are consistent with the agreed establishment (Anglophone countries) and directly with line departments (Francophone countries), to ensure that sufficient resources (and no more) are budgeted for the department’s approved establishment. During budget execution, personal emolumentsare identified and protected so that savings elsewhere in a department’s budget cannot be switched to meet payroll costs, and thereby evade establishment controls designed to maintain employment levels within set limits.

In the traditional establishment control model, the distinction between established and non-established positions is fundamental. Departments and agencies may fill established positions that are empty, but may not add positions to the establishment. Departments that wish to increase their establishments apply to the central personnel office or equivalent, and the latter then assesses the increased workload or the implications of a new policy mandate, and adjusts the department’s establishment accordingly. This is done in conjunction with the finance ministry so that additional budget resources to pay the higher costs of personal emoluments accompany any increase in positions.

The payroll must be closely linked to the establishment lists, the latter typically being the responsibility of the head of the civil service and the former the Ministry of Finance. Integrated personnel and payroll systems are now being introduced in some countries as a priority in restructuring processes. Staff added to the payroll must have been hired or appointed to a vacancy in a department’s approved establishment.

These arrangements generally apply to civilian central government and, separately, to subnational government. However, they do not apply to "non-established positions," such as those held by temporary workers. This can make the appointment of temporary workers a tempting option for hard-pressed managers seeking more personnel resources within tight establishment constraints. Equally, they do not apply to state-owned enterprises or to the armed forces. Click HERE to see a diagram that shows these key components of public employment.

Critiquesof the traditional model                                                           

Responsibility for the growth in civil service numbers in some developing countries has frequently been attributed to the mechanisms for establishment control. (See the page on Aggregate Employment & Wage Bill Concerns). The argument has been put that this traditional model places the line manager in permanent battle with the central personnel office over the size of the establishment, and that the line department faces no disincentives for seeking increases in establishment size through exaggerating their tasks and responsibilities.

This argument is not completely without merit, but it distracts attention from the larger driver of increases in total employment – the expansionist public policy agenda of the 1970s and 1980s. The paradigm of the interventionist state, the uncontrolled growth of policy and programs, and (for poorer countries) the rapid growth of aid aimed at promoting donor-favored projects and programs must take most of the responsibility for increased public sector employment. Arguably, weaknesses in establishment control arrangements and over-zealous hiring by line departments were a relatively minor part of the story.

There is however a further critique of the traditional model that can be sustained more readily. The traditional model is fundamentally incrementalist in nature. It encourages a focus on the marginal request, rather than staffing as a whole, mirroring the same bias in traditional budgeting arrangements. In the traditional model, the existing departmental establishment is treated as base, and the debate focuses on whether the request for additional posts is justified. Given the costs of amassing sufficient reliable information to review all programs on a regular basis, this is not an irrational strategy. However, by making the central personnel office or equivalent the arbiter of who and how many should be employed, the system constrains unit managers from employing the skills they deem necessary to carry out the job. Rigidity is imposed on the resources that managers are supposed to manage, as the model is administration not management.

However, moving into a more performance-oriented mode is not a step to be taken lightly as it rests on some complex preconditions and brings some risks. These are considered in more detail below.

Recentchanges in establishment control                                                

Many OECD countries seeking to move beyond the limitations of the traditional model and improve the efficiency and effectiveness of government have adopted a more managerialist approach to personnel management. Some have moved to a devolved "running cost" control system, in which managers are allowed to choose their own mix of staff and other resources to deliver programs efficiently and effectively within a hard budget constraint. A key goal of these changes is to allow managers to employ the skills they deem necessary to carry out the job, without the rigidities created by a central agency determining who and how many should be employed.

There are several preconditions for effective "running cost" control systems. These include:

  • Effective funds control framework. Regular, accurate financial reporting must be in place to assure large personnel commitments do not occur outside the approved budget. Further, the budget and accounting system must be in sufficient detail to allow separation of running form program costs, to detect any inappropriate subsidy of running costs from program funds. (See IMF page on Open Budget Preparation, Execution, and Reporting.)
  • Performance monitoring. Some rudimentary means of monitoring program output or quality should be in place. Managers have the freedom to ‘develop’ their own production function, but central Ministries should expect no reduction in output or service from the additional flexibility. (For more information see Regulation of Performance on the Public Expenditure website)
  • Performance culture. Civil service personnel, especially managers, should have internalized values for responsible financial management, mission achievement, economy and efficiency. This allows the funds control and performance monitoring systems to operate ex post for accountability, rather than ex ante for control.
  • Motive. Motive is provided by expectations of managers from the public and from political principals that services will continue to improve in quality or efficiency year on year. To a significant degree these expectations are based on past performance – in this sense, success in using the traditional control system partly laid the foundations for a running cost approach.
  • Opportunity. Opportunity is provided by robust budgetary and staffing information that allows detailed tracking of resource use. Without motive or opportunity, such devolution can be a recipe for unconstrained hiring or for over-compensation. Even within the OECD, at times of stress traditional establishment control, in the form of overall staffing caps or central controls on departmental structures, has occasionally reasserted itself over "running cost" control regimes.

It is important to note that these OECD "running cost" control regimes were built on the foundations of well-functioning traditional establishment control arrangements.

Whatdo we mean by "pay"?                                                                  

Base wages and salaries are just one element of the total rewards received by civil servants. The 9 elements of total rewards can be seen in this table. For further details, see the page on  rewards and incentives. Please note that the examples noted in each cell are not exhaustive. The relative importance of any cell to a public employee’s total rewards can (and does) vary dramatically between countries. Note, also, that allowances may still be discretionary, even if based in contracts.



non-contractual/ intangible



current rewards

base rewards

1. base wage/salary

2. health insurance

3. job security, prestige, social privileges


4. transportation, housing, meals, telephone, travel, cost-of-living

5. transportation, housing, meals, travel

6. trips abroad, training

future expectations

7. pension

8. housing, land, etc.

9.reputation,re-employment after retirement

This website uses the following conventions:

  • personal emoluments = cells 1 and 4 (current monetary rewards and allowances)
  • personal disposable income = personal emoluments minus employer deductions for income tax withholding, provident fund/pension contributions, etc.
  • total compensation = cells 1, 2, 4 and 5 (contractually provided current rewards and allowances)
  • total rewards = cells 1 to 9 inclusive (contractual and non-contractual, current and expected rewards and allowances)

Problemsof pay policy                                                                          

Often a high proportion of the real earnings of civil servants is not monetized; many benefits are not captured by the "personal emoluments" line item in the budget. For example, permanent secretaries in Zambia appear to earn only 5 or 6 times as much as the lowest paid government employees. However, when the in-kind benefits such as housing, cars, telephones, etc. are taken into account, permanent secretaries earn as much as 50 times as much as the lowest paid. The problem from the point of view of the permanent secretary is that only a small proportion of his or her total benefits is personal disposable income.

At the aggregate level, a number of developing countries continue to grapple with the dilemma of reducing the overall public sector wage bill while simultaneously trying to improve remuneration to attract or retain skilled staff. Their challenge is that civil service pay is subject to many diverse policy pressures. In many developing countries, civil service pay policy in the 1960s reflected the concern that governments should be a wage leader, setting salary scales that would serve as a model for the modern private sector. Low end wage rates were determined nationally by independent commissions and wages councils, to respond to policy concerns about minimum wages. Further distortions arose from the political commitment to employ school leavers and the "bonding" of graduates to the public sector for a minimum period. As policy overload began to appear and staff numbers continued to expand despite static revenues, those few pay adjustments for inflation that were possible began to be biased towards the low end, largely for political reasons. To compensate the top end there were extra allowances and deliberately less visible non-wage benefits.

For further discussion of public service pay policy in Anglophone Africa, click here.

Recentchanges in pay policy                                                                 

Within the OECD many governments have resolved to offer more competitive wages and reward performance. Their key task has been to encourage innovation and a greater results focus in an essentially well-skilled public sector work force. This has led to the introduction of some flexibilities in pay in the OECD. Still, in most OECD countries base wages continue to average more than 90% of civil servants’ total wage package and there is strikingly little evidence that performance-related pay arrangements contribute to improved performance. See the materials in this site on pay flexibility and on individual performance managementfor further details.

The challenge facing many developing countries is dramatically different. There are, for example, reportedly more Ghanaian doctors, architects and engineers in the city of Toronto than in Accra. What remuneration would be necessary to attract them back or prevent others from leaving? This means pay levels that are sufficient for government to recruit, motivate and retain technical, professional and managerial staff within the country in competition with local modern sector and international employers. Politically this is challenging, since it implies radical decompression of formal scales.

Sequencingin the reform of establishment control and pay systems     

It is important to establish precisely the nature of the performance problem that is to be addressed. As noted above, OECD countries have been addressing efficiency concerns, not over-hiring, with their reforms in establishment control and the shift towards "running cost" regimes. If over-hiring and unaffordable wage bills are the primary objects of reform, then the first order concern is to reduce the policy ambitions of government. This means reforms in the policy management arrangements within government, so that the costs of new policy proposals can be compared to the cost of existing policies. See the materials in this site on poor responsiveness to changing priorities.

To the extent that efficiency is a concern in addition to policy overload, then the information and performance management systems within many developing countries are insufficiently robust to risk radical experiments with devolved "running cost" control systems. In many settings, civil service managers do not have up-to-date lists of who they employ to do what. It is a fertile environment for ghost workers. Departments may also ignore establishment limits, and without effective collaboration between the personnel and finance ministries in setting budgets, finance ministries often have provided the resources for hiring beyond establishments. In systems such as these, it is often necessary to make the rebuilding of traditional establishment control systems the first step in a reform program. (See the page on Civil Service Censuses.)

When traditional control systems are working better, and formal and informal rules are better aligned, a cautious move toward a more "managerial" establishment control model may be warranted. The tests for such devolved systems are:

1. Can the budget constraints really be made hard?

If the implemented budget is to be the enacted budget, then there must be formal procedures for requesting modifications during the year. These requests must be restricted to those made in the hope of being able to expend resources, not ex post to finance already incurred expenses. In cases where there is a conflict between current law (e.g. hire X thousand more teachers in fiscal year Y) and the budget (the budget for fiscal year Y only supports ½ X thousand teachers), the budget must control actual hiring.

2. Are there adequate operational spending control mechanisms?

Administrative mechanisms must be in place (at the center of government and in each agency) to monitor and control spending during the year, and rectify spending imbalances. In the case of personnel, this control may operate ex ante through pre-certification by the finance officer that funds are available for a particular hire. Or controls can operate ex post through internal and/or external audit to detect overspending, determine the cause, and assess a penalty. Further, budget and accounting recording and reporting systems need to contain sufficient detail to separate running costs from program costs to detect where program funds are being inappropriately used to support running costs.

3. Can responsibilities be completely aligned?

A central tenet of organizational or system design in devolved systems is to align management authority with financial accountability. Those who make the spending decisions must also bear the cost of the decisions; those held responsible for financing must also have the authority to control or influence spending. Misaligned responsibilities within devolved systems create opportunities for moral hazard and over-consumption.

In combining improvements in establishment control systems and pay policy, the challenge facing many developing countries is to bridge two separate models. One is driven largely by donors’ concerns for the execution of their projects, and includes enclaving, salary supplements, and the ring fencing of critical areas to escape dysfunctional public sector pay scales. The other model reflects an interest in more systemic civil service reform, including improved wage bill control, monetization of job benefits, and removal of "ghost workers."

Monetizing all tangible benefits is often identified as one of the first steps in pay reform. Certainly there is every reason to expose the total compensation of different categories of staff for planning purposes. However, monetizing carries some risks. Non-transparent compensation is sub-optimal but is ultimately a coping mechanism. Although it is inefficient and inequitable, increasing transparency prematurely is likely to generate uncontrollable and unaffordable pressure for equalization of benefits and across the board pay increases. Similarly, the relevance of performance-related pay should perhaps be resisted or at least questioned.

The first step in a pay reform strategy is to determine the medium-term resources envelope. This will indicate how much headroom there is for a pay reform, or, conversely, the extent to which government is overcommitted in terms of staffing levels. Since the arithmetic generally demonstrates that revenues are insufficient to improve incentives across the board, wage increases often must be focused on key labor categories, usually those for which current salary levels are well below market (e.g., managerial, technical and professional staff). It is important, however, to consider the total reward package before making these decisions. Some public employees are paid (or benefit) more than most people realize.


This page was developed by Harry Garnettand Mike Stevensof the World Bank, with contributions from Bill Dorotinsky, Gary Reid, Jeffrey Rinneand Nick Manning. It was submitted on 12/7/00.


Permanent URL for this page: