VII. Public Sector Pay
- Ke-young Chu, and Richard Hemming
- Published Date:
- September 1991
What determines wages and salaries in the public sector?
Is public sector pay out of line with that in comparable private sector jobs? Is this desirable?
How important are fringe benefits? Are they necessary?
What measures can be taken to control the growth of the wage and salary bill? Do these measures have unintended consequences?
Public Sector Wage and Salary Determination
Public sector wages are subject to different forces than private sector wages. If the salary paid to certain occupational groups in the civil service is substantially lower than it is in the private sector, it would be expected that the civil service would have difficulty hiring labor, but the government would not be subject to the same kinds of financial pressures to fill vacancies as private sector managers. Similarly, if civil service salaries were higher than they needed to be to hire suitably qualified candidates, the government would not suffer the same kind of financial consequences as a private sector enterprise functioning in a competitive environment. Public sector pay awards in many countries are also undoubtedly affected by the relatively high degree of unionization in the public sector. As a result, civil service pay may diverge from private sector pay for long periods of time unless pay-setting mechanisms are adopted to ensure comparability.
The differences in the manner in which public and private sector wages are set should not, however, be exaggerated. A position of market dominance can partially insulate some private sector enterprises from market forces, and give them a degree of monopsony power in hiring labor. At the same time, some occupational groups are clearly able to exploit a position of monopoly power in the sale of their services, thereby extracting a rent over and above the wages they would need to be paid for the same supply of services in a competitive market.
Pay Disequilibria in the Public Sector
There are obvious costs to society in setting public sector wages at unreasonably high levels, unless paying civil service employees at levels above their reservation wage is a means of achieving a desired redistribution of income. In addition to their budgetary costs, high wage settlements in the public sector can influence settlements throughout the economy. Although disentangling the causal link between wage settlements can be difficult—does one high nominal wage settlement lead to another, or are they both a reflection of high inflationary expectations?—the presence of parallelism in wage settlements implies that public sector pay policy can have a macroeconomic impact that reflects more than its direct expenditure consequences.
Are wages and salaries in the public sector too high, in the sense that they could be significantly less but still be high enough to attract and retain a suitably qualified work force? This is not a question that can be satisfactorily answered by reference to aggregate indicators of compensation or income in the public and private sectors, although these kinds of indicators allow some interesting international comparisons. In particular, public sector salaries in developing countries are typically much higher in relation to per capita income than are public sector salaries in the industrialized countries. Many different influences could contribute to these differences. Even these aggregate statistics, however, are plagued by problems of data comparability, and in any case a satisfactory answer to the question posed above would require a wealth of information for each distinct occupational group and each country under study.
There is some evidence in a number of African countries, and rather anecdotal evidence in other developing countries, that average public wages and salaries may be too low. In Uganda, for example, in the period 1975-83, real salaries fell by between
The response to declining real salaries seems to have been a combination of increasing corruption, moonlighting, absenteeism, declining performance standards, and reductions in the supply of public services, often in critical areas such as education and health. Ironically, while the wage bill was accounting for an excessive share of public expenditure because of the swollen ranks of the civil service, per capita salaries in some of these countries were simply too low to compensate adequately a well-trained and well-motivated civil service. The appropriate policy response to this situation is to reduce employment and increase average levels of compensation. Inevitably, this policy would encounter strong resistance from those civil servants who would be dropped from the payrolls.
In addition to reducing civil-service-wide real salaries, efforts at restraining the growth of the wage bill in the civil service in many developing countries have contributed to a substantial compression of the pay structure. Although part of the decline that has been observed in the relative pay rate of the more highly skilled and the professional occupational groupings in such countries as Jamaica, Peru, and some of the African countries noted above could be justified on the ground that investment in education has reduced the relative scarcity of the better educated, this phenomenon has also been associated with shortages in the highly skilled occupations. Because pay scales in public enterprises are generally not subject to the same pressures as civil service pay scales, senior executives are lost to both the private sector and to public enterprises. It is also alleged that pay compression leads to an increase in other forms of compensation, which, being less visible, may not be as politically sensitive or as controllable as wages and salaries.
While pay differentials cannot in general be fixed by the market, it would not be efficient to let them be determined by line managers, because the result would be different levels of compensation within the civil service for the same jobs and levels of experience and expertise. A policy is necessary for the civil service as a whole and, to the extent possible, public sector wages and salaries should be linked to their nearest equivalents in the private sector. This principle of comparable worth is applied in a number of industrialized countries and involves an elaborate compilation of data on private sector wage and salary settlements for jobs deemed to be equivalent to the various occupations and grades found in the public sector. However, such a comparison can pose a problem for professions, such as tax collectors, where it can be difficult to determine the private sector equivalent. Although it may be hard to gauge when salaries for such a profession are out of line, excessively low salaries will be reflected in growing numbers of vacancies and in recruitment difficulties, while high salaries can result in a superabundance of well-qualified candidates.
Fringe benefits in the public sector come in all shapes and sizes. They include spouse and dependents allowances, pensions, health and disability insurance, free or subsidized housing, free or subsidized lunches, transportation allowances, and leave with pay. To this list can be added travel allowances to the extent that they exceed normal travel costs. In addition to fringe benefits proper, in many countries the base salary is supplemented by bonuses of various kinds, supplements for dangerous or unpleasant work, and the like.
The great variety of fringe benefits and the terms on which they are offered means that hard data on their importance is typically lacking, but they are obviously important in the upper echelons of the civil service in many countries, and they can be important at lower levels as well. To refer again to the African experience, nonsalary compensation in Senegal, which mainly took the form of allowances (cash payments), represented about 45 percent of base salary in December 1984. Total fringe benefits in Nigeria have been estimated to represent from 35 percent to 100 percent of base salary.
Impact and policies
Fringe benefits may be classified according to four criteria: (i) eligibility (across-the-board or restricted in some way), (ii) form (in cash or in kind), (iii) timing (deferred or contemporaneous with employment), and (iv) relation with base pay (related or unrelated). Thus, a pension is a deferred cash benefit, to which eligibility restrictions may or may not apply and which may or may not be related to base pay. To take another example, free housing for senior civil servants is a restricted benefit in kind that is contemporaneous with employment but which may bear little relation to base pay.
Fringe benefits obviously have a role to play in a well-designed compensation policy, but certain kinds of fringe benefits are also highly inefficient means of compensation that can create perverse incentive effects. For example, the value of free housing provided to one civil servant may be substantially different from the value of the housing provided another, and in each case the amount of housing supplied may be unrelated to their needs. If housing is in short supply, then a substantial real income gap is created between those lucky enough to enjoy this benefit and their less fortunate colleagues, and this gap may be unrelated to differences in their rank within the civil service. Moreover, the provision of subsidized or free housing creates a substantial disincentive to regional and occupational mobility. Excessive travel allowances obviously also stimulate excessive travel, and so forth.
As a general rule of policy, benefits in kind or subsidies tied to the consumption of one particular good are undesirable on efficiency grounds, because they restrict the freedom to substitute consumption of one good or service for another. In the case of free provision of housing or transportation, moreover, what may be an important part of the compensation package is not included in the budgetary estimate of civil service compensation or under any other budgetary heading, so that the budgetary presentation suffers from a lack of transparency. An exception to this rule is justifiable for subsidized meals, because these can provide a strong incentive for a more productive work day as well as a boost to productivity in low-income countries by improving nutrition. But in the case of most benefits in kind, it is more efficient to replace them with a compensating adjustment in pay and allow civil servants to pay for the quantity of the goods or services they consume. Policy reforms along these lines were adopted in Indonesia, Liberia and other countries, where official automobiles previously given to civil servants were sold to them, and in Botswana, where civil servants were encouraged to buy or rent the government-owned housing.
Although fringe benefits do not necessarily have to be related to pay—i.e., they can be a flat rate, for example in the form of a flat rate dependency or transportation allowance—if a substantial share of the total compensation package is unrelated to base salary, then the motivational role of merit awards and promotions is reduced. A proliferation of flat rate benefits also makes the calculation and comparison of pay relativities within the civil service more difficult, as well as complicating indexation policy.
From the point of view of expenditure policy, the most important characteristics of fringe benefits are their relative importance in the overall compensation package, their relationship with base pay, and the degree to which they can be subjected to the same controls as base salary. Benefits in kind or benefits tied to the consumption of particular goods are in general undesirable and should, where possible, be replaced by cash substitutes.
Wage and Salary Restraint
A basic consideration of any program to restrain the real wages of the civil service relative to the private sector must be the impact of such a policy on morale and productivity within the civil service. Are there strong grounds for thinking that public sector salaries are too high, and that they could be reduced substantially without prompting an exodus to the private sector, increased absenteeism, or reduced productivity and permanent damage to civil service morale? If this is not the case, a policy of wage restraint may achieve the short-run goal of reducing public expenditure but lead to an undesirable reduction in the quality and quantity of public services. If public salaries are already low, further reductions can threaten the provision of basic public services.
As noted above, per capita real compensation had declined substantially in some African countries, and there was impressionistic evidence that these declines caused reductions in work effort if not in employment. Nonetheless, reductions in real wages by means of a wage freeze or a policy of less than full indexation to price changes will reduce real expenditure. Such policies may be less difficult politically than policies to reduce employment, even if the real problem with the public sector wage bill is not the level of real wages but the excessive size of the payroll. Nonetheless, such a policy can be vitiated if there is no control over the hiring process. The substitution of fringe benefits for wages and salaries can also undermine efforts to contain pay. Moreover, the indiscriminate application of across-the-board restraint will often result in serious sectoral manpower shortages.
Policies affecting advancement within and between ranges or grades also have important consequences for wage policy. In a number of countries, efforts to freeze wages were thwarted by increases in either the average rate of advancement within grades or in the frequency of promotions, indicating that there was a serious lack of control of the overall salary structure. Such wage drift may simply not be foreseen, and cannot be contained unless there is centralized control over the granting of merit or longevity increments and promotions, and a careful analysis of the cost of alternative policies. Although all the details of compensation policy cannot be dealt with adequately in a stabilization or structural adjustment program, it is important to estimate the share of average wage increases accounted for by wage drift and to have an idea of the scope that uncontrolled wage drift offers for offsetting the intended impact of wage restraint on the average salary of civil servants.
Public sector employment and pay in Argentina
The recent evolution of public sector employment and pay in Argentina illustrates the tendency for relatively rapid growth of employment in the public sector coupled with a decline in average real wages and salaries. However, data on public sector pay and employment in Argentina, although better than those of many countries, are nonetheless deficient: in particular, basic pay can exclude a substantial part of total compensation. As a consequence, any interpretation of the data has to be made with some caution.
Public sector employment in Argentina has accounted for about 18 percent of estimated total employment in recent years. Of some 1,960,000 public sector employees in 1988, regional- and local-government employees accounted for almost 50 percent, reflecting in part the important role the provinces play in the provision of social services. Employment at the National Administration (central government) level accounted for 34 percent, while employment in public enterprises accounted for 18 percent. The share of employment in the public sector in Argentina is high in comparison to other countries with a comparable per capita GDP and tax base.
The number of public sector employees in Argentina is estimated to have grown at an average annual rate of about 2.2 percent in the 1980-88 period, with the fastest-growing component being employment at the regional level. Employment in the private sector is estimated to have increased at about 1.2 percent annually over the same period. The relatively rapid growth of public employment at a time when the growth of GDP was essentially stagnant contributed to deteriorating public finances and led to pressure to restrain public sector pay. As a result of these pressures, the average real wage of public sector employees fell relative to that in the private sector. While private sector real manufacturing wages rose by 12 percent between 1982 and 1988, the real basic wage of middle-ranking civil servants fell by 17 percent. Moreover, this relative decline continued when private sector wages were themselves subsequently depressed by rising unemployment. In 1989 private sector wages fell to slightly below their 1982 average, but public sector real wages fell even more sharply. In addition to the decline in average real wages of civil servants, there is evidence that the compression of the public sector salary scale increased during the 1980s. The decline in real wages also led to an increase in the number of civil servants holding a second job, and public sector productivity has suffered as a result.
A policy to reduce public sector employment in the National Administration and the public enterprises has now been adopted. In addition to the decline in public sector employment that will result from the privatization of a number of public enterprises, and the elimination or curtailment in the activities of a number of public financial institutions, employment reductions are being sought through the implementation of early retirement, a stricter interpretation of the statutory retirement provisions, and layoffs in branches of government departments deemed to be overstaffed. Once employment levels have been reduced sufficiently, it will be possible to increase real public sector pay, and to widen relativities through discretionary increases for those higher occupational groups whose remuneration has fallen particularly sharply relative to comparable groups in the private sector.
Public sector employment and pay in Ghana
The 1970s and early 1980s were characterized by rapid growth in public employment in Ghana, and dramatic declines in productivity and real wages. By 1983, real monthly earnings per government employee were less than 11 percent of their 1975 level. The result was an overstaffed, poorly trained and unmotivated civil service. The ongoing civil service reform program, initiated in 1987, has three primary goals. First, the Government aims to reduce overstaffing, mostly at lower levels, while at the same time relieving shortages in high-skilled areas. Second, to bring this about, it has become necessary to improve the competitiveness of civil service pay, particularly at higher levels; this requires both a rise in overall public sector wages as well as increasing pay relativities between highest- and lowest-paid civil servants. Finally, the reform is attempting to increase productivity by providing incentives for performance, restructuring the job grading system, and enhancing training and management.
Government employment—including the civil service, the Ghana Education Service, and the prisons, fire and foreign services— totaled 305,000 in December 1986, of which 144,000 were in the civil service. By August 1990, total public employment had been reduced to 274,000, with 112,000 in the civil service. A “Special Efficiency” budget has been established to provide severance pay and retraining/relocating grants for retrenched public employees. A significant widening of pay relativities has also been achieved. The ratio of total compensation for highest- to lowest-paid civil servant was increased from 5.4:1 to 9.4:1 between 1987 and 1990. The goal, in the context of the World Bank Structural Adjustment Program, is a ratio of 13.4:1. The authorities plan to introduce performance-related pay in 1992. In addition, the possibility of privatizing some low-skilled civil service activities, such as cleaning and maintenance, is being considered.
Lindauer David L. “Government Wage and Employment Policy: an Analytical Treatment with Special Reference to Africa,” Discussion Paper No. 1987-3 (Washington: World BankMarch 1987).
Lindauer DavidL. Meesook Oey Astra and Parita Suebsaeng “Government Wage Policy in Africa—Some Findings and Policy Issues,” World Bank Research Observer (Washington: World BankJanuary 1988).
Nunberg Barbara “Public Sector Pay and Employment Reform,” Country Economic Department Working Paper Series 113 (Washington: World BankOctober 1988).
Schiller Christian “Government Pay Policies and Structural Adjustment,” Working Paper 88/73 (Washington: International Monetary Fund1988).