Chapter

III Issues in the Analysis of Public Sector Employment and Wages: Leverage Implications of Public Employment

Author(s):
Alan Tait, and Peter Heller
Published Date:
October 1983
Share
  • ShareShare
Show Summary Details

A principal motive for analyzing the size of government is the belief that government employment and wage policies have critical implications for wage determination throughout the economy. The larger the government share of employment, the more likely it is to dominate wage rates and awards not only for public sector employees but for the private sector as well, and thus to have a sigificant degree of “leverage.” What is the fulcrum point at which government decisions on employment and wages affect employment and pay throughout the economy?

This sort of question is important because in both industrial and developing countries employees in the public sector can view their employer as having no limit to the financial resources available for wages; they start to view themselves as having access to the money supply.5 Their success in claiming wages higher than their productivity would merit exacerbates the contrast with the private sector (who, eventually, must finance the higher public sector pay). Rapid and unexpected increases in public sector wages have undermined macroeconomic stabilization policies and Fund programs,6 and such problems can extend even to queries about municipal fiscal integrity.7

Measures of the Size of Government Employment

Most studies on the size of government tend to focus on the level of government expenditure or revenue and its relationship to GDP. Another equally germane measure would be the magnitude of government employment. The absolute employment number provides a measure of the quantum of input involved in the provision of public services. The relationships of these employment numbers to both population and measures of the labor force employed in the nonagricultural sector are likely to be indicative of the impact of public sector wage policies on wage rates in the economy, the distribution of income, and the structure of output in the economy. As governments have little direct influence on agricultural wage rates (unless common minimum wage rates apply to the agricultural sector), it is the size of government employment relative to total nonagricultural employment that has been taken as the measure of potential leverage.

Public sector employment may occur at the central government level, state and local authority level, and in the nonfinancial public enterprise sector. General government is defined to include both central government and state and local government employment; public sector employment combines central, state, and local governments, and the nonfinancial public enterprise (NPE) sector. In Appendix I, Tables 20-22 provide the complete set of data on the absolute size of government employment as well as their relationship to employment in the nonagricultural sector, as reflected in ILO statistics,8 and to the total population. Table 1 summarizes the means of the different variables, classified by region.

Table 1.Share of Government Employment in Total Nonagricultural Sector Employment and Population, by Level of Government and Region: Means and Standard Deviations1
Developing Countries
OECD Industrial CountriesTotal sample of countriesAfricaAsiaLatin America
(As a share of nonagricultural employment, in percent)
Central government(X)8.723.430.813.920.7
(s)(5.7)(16.2)(15.0)(3.9)(21.1)
(n)16311359
State and local government(X)11.64.02.18.04.2
(X)(6.0)(7.3)(2.6)(14.7)(4.4)
(n)163516510
General government(X)19.226.033.022.220.7
(s)(5.9)(13.7)(15.1)(16.3)(4.6)
(n)21311367
Nonfinancial public enterprises(X)4.113.918.715.75.5
(s)(2.7)(11.9)(14.2)(10.8)(4.2)
(n)1418845
Public sector employment(X)24.243.954.436.027.4
(s)(7.8)(22.1)(21.2)(23.1)(13.2)
(n)14231255
(As a share of total population; per 100 inhabitants)
Central government(X)3.12.41.82.62.6
(s)(1.6)(1.5)(1.3)(1.8)(1.2)
(n)163516510
State and local government(X)4.60.40.20.40.8
(s)(3.0)(0.7)(0.2)(0.4)(1.1)
(n)16311557
General government(X)7.23.01.93.14.6
(s)(2.7)(2.0)(1.3)(1.3)(2.8)
(n)21361768
Nonfinancial public enterprises(X)1.51.10.82.10.9
(s)(0.9)(1.2)(0.7)(2.2)(0.8)
(n)14201045
Public sector employment(X)9.03.72.94.64.8
(s)(3.3)(2.2)(2.1)(2.5)(2.0)
(n)14201144
Sources: Tables 21 and 22.

X = mean; s = standard deviation; n = number of observations in the sample.

Sources: Tables 21 and 22.

X = mean; s = standard deviation; n = number of observations in the sample.

In contrasting the relative importance of government employment in the industrial OECD countries (Australia, Austria, Belgium, Canada, Denmark, Finland, France, the Federal Republic of Germany, Iceland, Ireland, Italy, Japan, Luxembourg, the Netherlands, New Zealand, Norway, Spain, Sweden, Switzerland, the United Kingdom, and the United States) and the developing world, certain patterns emerge clearly. First, central government is far more significant in nonagricultural sector employment in the developing countries, averaging 23 percent of such employment in contrast to only 9 percent in the OECD countries. This fact reflects primarily the small share of nonagricultural sector employment in the total labor force in developing countries. In some African countries, the role of the central government in nonagricultural employment is striking, reaching up to 42-46 percent in Benin, Tanzania, and Zambia, and averaging 31 percent for the African countries in the sample. By contrast, in the OECD countries, nonagricultural sector employment is a much larger percentage of the total labor force, such that the role of central government employment on employment in the nonagricultural sector is much less; nevertheless, it is noteworthy that in New Zealand 24 percent of its nonagricultural sector is employed in the central government and in Belgium, 18 percent. The lower figures—3 percent for the Federal Republic of Germany and Canada, and 4.5 percent for the United States—are represented by the federal countries, while the more typical figures for a centrally governed economy in the OECD might be about 10 percent (the United Kingdom) or 8 percent (the Netherlands).9

There is considerably larger variance in the ratio of central government employment to nonagricultural sector employment in the developing countries than in the OECD countries. In the developing countries one standard deviation from the mean of this ratio implies a range from 7 percent to 50 percent. In the OECD countries, the range would only be 3 percent to 14 percent.

The clear message from these statistics is the significant impact that government policy on wages and salaries is likely to have on the overall remuneration of employees in the nonagricultural sector in developing countries. Without even taking into account state, local, and nonfinancial public enterprise employment, central government decisions on wages and salaries in developing countries are likely to affect from 15 to 40 percent of the urban labor market, and therefore to have a pervasive effect on domestic unit wage costs. Clearly, in terms of formal incomes policies or in general wage bargaining, this is likely to be an important influence.

An alternative perspective on central government employment is suggested by an examination of the number of employees per capita. This is a rough measure of the magnitude of public services provided at this level of government, although it, of course, says nothing about the quality of the services or the efficiency with which they are provided. Using this measure, the number of central government employees per capita is considerably higher in industrial countries than in developing countries; this is so despite the relatively greater importance of state and local government employment in the OECD countries. For example, as a share of the population, central government employment in OECD countries averages 3.1 percent as opposed to 2.4 percent in developing countries, with the range of developing countries spanning only 1.8 percent for Africa to 2.6 percent for Asia and Latin America. Some of the countries with a striking number of central government employees per capita are New Zealand (6.9 per 100 inhabitants), Belgium (4.9), Iceland (5.5), Mauritius (5.4), and Swaziland (3.4).

Of course, the legitimate point can be made that it is unreasonable to look only at central government employment. particularly when federal constitutions are considered. The influence of local authorities on wage rate determination in their locality can be pervasive; rates set for unskilled labor (for example, garbage collection, road maintenance), or for skilled local services (for example, teachers, librarians, and administrators) can set a standard for local private sector employers to match. The countries with the largest share of state and local government employees in nonagricultural sector employment were the federal countries, principally within the OECD group, the United States (14.3 percent), Australia (19.5 percent), and among developing countries, India (34.2 percent). However, there are some interesting anomalies particularly within the OECD, where Denmark (19.6 percent), Sweden (25.1 percent), and the United Kingdom (12.6 percent) have remarkably large local governments for countries that are often thought to be dominated by a unified central government.

As a group, state and local government employment averages almost 12 percent of nonagricultural sector employment in the OECD countries in contrast to 4 percent in the developing countries. The difference is even more dramatic if one calculates state and local government employees as a share of the total population: 4.6 per 100 inhabitants in the OECD countries, in contrast to 0.4 in the developing countries. Typically, the latter are much more centralized governments, with the share of local government in nonagricultural sector employment ranging from 2 percent in Africa to 8 percent in Asia. The mean employment share of the central government in total general government employment in those countries is approximately 85 percent (Table 2 and Appendix I, Table 23). This contrasts with a ratio of only 43 percent in the OECD countries. It seems probable that “leverage” will be more powerful for local authorities in OECD countries than in developing countries.

Table 2.Share of Different Units of Government in Total Employment of General Government and the Public Sector, by Level of Government and Region: Means and Standard Deviations1
Developing Countries
OECD Industrial CountriesTotal sample of countriesAfricaAsiaLatin America
(As a share of general government employment; in percent)
Central government(X)42.485.490.279.381.4
(s)(22.2)(18.6)(10.6)(24.5)(21.2)
State and local government(X)57.614.69.820.718.6
(s)(23.7)(19.2)(9.2)(24.4)(22.1)
Number of cases16311557
(As a share of public sector employment; in percent)
Central government(X)34.958.765.543.365.0
(s)(19.1)(20.7)(15.9)(20.2)(22.8)
State and local government(X)48.612.45.517.715.6
(s)(22.3)(16.2)(6.5)(20.1)(20.0)
General government(X)83.571.171.061.080.6
(s)(8.3)(15.0(12.3)(22.4)(12.4)
Nonfinancial public enterprises(X)16.528.929.039.019.4
(s)(8.3)(15.5)(12.9)(22.4)(12.4)
Number of cases14191044
Source: Table 23.

X = mean; s = standard deviation.

Source: Table 23.

X = mean; s = standard deviation.

Aggregating employees at the central, state, and local governmental levels, the number of general government employees per capita proves to be significantly more important in the developed countries; the OECD countries average more than 7 per 100 inhabitants in contrast to 3 in the developing countries. Among OECD countries, it is precisely those countries with the centralized government combined with a remarkably large local government component that have the highest number of general government employees per capita, for example, Sweden (14.7), Denmark (11.4), and the United Kingdom (9.6) (Appendix I, Table 21). There also appear to be some regional variations in the significance of general government employment, although the comparisons are more limited given the paucity of data on local authority employment in many countries. The share of general government employees in nonagricultural sector employment is significantly larger in Africa and Asia than in Latin America. The reverse relationship holds with respect to the number of employees on a per capita basis, averaging 3.1 employees per 100 inhabitants in Asia and 4.6 in Latin America, in contrast to only 1.9 in the African countries in the sample.

The broadest definition of public sector employment used here embodies central government, state, local, and nonfinancial public enterprise employment. The employees of nonfinancial public enterprises loom much larger in developing countries, averaging 14 percent of nonagricultural sector employment in contrast to only 4 percent in the OECD countries. The share of nonfinancial public enterprise employees appears to be significantly higher in Africa than in Latin America or, with respect to the mean for the developing world as a whole, equaling 19 percent in the African region, although this is clearly based on a limited sample of countries.

The role of the nonfinancial public enterprise sector in the public sector is also considerably larger in developing countries; as a share of total public sector employment, it averages 29 percent in the developing countries. In the OECD countries, this ratio equals only 16.5 percent, with the remaining government employees divided between the central (35 percent) and state and local (49 percent) government levels. If one argues that the more centralized the employment the easier it is to impose a common wage policy, one would assume that developing countries have greater leverage on general government wage policy and somewhat less influence over wage rates in the nonfinancial public enterprise sector. Nevertheless, there is evidence that in some of those countries (for example, Zambia and Zaïre) and in some industrial countries (for example, the United Kingdom and France) wage awards to workers in public industries are viewed as crucial for wage determination countrywide.10

The more striking figures here are that public sector employees average 44 percent of nonagricultural sector employment in developing countries and 24 percent for OECD countries. Among developing countries, the share of public sector employees in nonagricultural employment in Africa reaches 54.4 percent, in contrast to 27.4 percent in Latin America and 36.0 percent in Asia. In some developing countries, the ratio reaches as high as 87 percent—in Benin—followed closely by Ghana (74 percent), Zambia (81 percent), and India (72 percent) (Appendix I, Table 22). Within the OECD, it is as noteworthy that, in New Zealand, 36.5 percent of nonagricultural sector employment is in the public sector, as it is that this ratio reaches 26.8 percent in Ireland, 33.8 percent in Sweden, 32.8 percent in Belgium, and 30.8 percent in the United Kingdom. The gap between the “least governed” economy (the public sector in the United States employs 19.6 percent of nonagricultural sector employed) and the “most governed” economies (New Zealand and Sweden) is large.

On a per capita basis, the ratio of public sector employees in the OECD countries is almost double that observed in this sample of developing countries. The ratio is higher in Latin America—public sector employees averaging 4.8 per 100 inhabitants in contrast to 2.9 in Africa, although again this is based on a limited sample of countries.

The literature suggests that, whereas the influence of public sector wage awards is important in many European countries (for example, the United Kingdom, Sweden, France, and Ireland), it is not in the United States. The crucial “breakpoint” might be said to be public sector employment between 20-25 percent of the nonagricultural sector employed. Over that figure, public wage awards seem to affect national wage rate determination; under it, the leverage appears to be less important.11

Measures of the Size of Government: Wages and Salaries

The impact of government employment on an economy can also be examined in terms of the weight of government wages in total output and value added. There are several ways of viewing this relationship. Table 3 shows the mean share of wages for central government, state and local government, and nonfinancial public enterprises, respectively, in general government and public sector wages. (See also Appendix I, Table 24.) As one would expect, federal governments typically exhibit a structure where more than 70 percent of total government wages are paid to state and local governments (for example, in the United States. Canada, and the Federal Republic of Germany). The only country in the sample outside the OECD with a similar government structure (Brazil) shows a 40-60 split between central and state and local government wages.

Table 3.Share of Different Units of Government in Total Payroll of General Government and the Public Sector: Means and Standard Deviations1
Developing Countries
OECD Industrial CountriesTotal sample of countriesAfricaAsiaLatin America
(As a share of general government wages; in percent)
Central government(X)45.591,296.780.9
(s)(22.4)(17.7)(5.9)(…)(24.9)
State and local government(X)54.58.83.319.1
(s)(22.4)(17.7)(5.9)(…)(24.9)
Number of cases1123929
(As a share of public sector wages; in percent)
Central government(X)43.664.075.954.4
(s)(23.9)(25.5)(17.7)(…)(28.4)
State and local government(X)38.914.22.423.7
(s)(22.8)(16.2)(3.8)(…)(16.3)
Nonfinancial public enterprises(X)17.521.821.721.9
(s)(10.9)(15.5)(16.4)(…)(16.7)
Number of cases59415
Source: Table 24.

X = mean; s = standard deviation.

Source: Table 24.

X = mean; s = standard deviation.

These figures are not unexpected, but what is interesting is how high the local government wage bill is, compared with that of the central government, in many countries where government is generally thought of as centrally dominated. For instance, it is striking that in Japan 69 percent of the wage bill is paid to local government officials, and almost 70 percent in Denmark. In the Netherlands, 58 percent is paid to local government and in the United Kingdom, 51 percent; in other countries (for example, Argentina and Costa Rica) the percentages paid to local government are still large (50 percent and 43 percent, respectively). This situation emphasizes how important wage settlements are at the local level in such countries and how important it is, in speaking of national wage policy, to appreciate whether or not the central government has de facto control over local government pay and hiring.12 For example, one of the major confrontations in the last four years in the United Kingdom has been between the central government attempting to enforce its pay policy down to the local level and local authorities resisting such pressures.

In terms of the total public sector wage bill, the United Kingdom is noteworthy for having an almost equal split of total wage and salary payments between the central government, local government, and the nonfinancial public enterprise sector. In other OECD countries, the nonfinancial public enterprises account for a smaller share of the total public sector wages bill. However, in some developing countries, the wage bill of the nonfinancial public enterprises sector can be as high as 50 percent of the total public sector wage bill (e.g., in Brazil) and 45 percent (e.g., in Zambia). Again, it is clear that in these countries public sector decisions on payments to employees in public enterprises have influence not only on the public sector’s wage bill but also on the wage determination process in the country as a whole, at least in the nonagricultural sector.

Table 4 shows central government wages as a percentage of total wages, national income at market prices, and GDP. There are interesting features in these figures; for instance, the highest proportion is that of Greece (19 percent of national income). Even countries with an extremely high proportion of public sector employees among the nonagriculturally employed do not necessarily have a particularly large claim on national income, for example, India at 3.8 percent and New Zealand at 13.3 percent (Appendix I, Table 25).

Table 4.Government Wages Relative to Total Wages, National Income, and GDP: Means and Standard Deviations1
Developing Countries
OECD Industrial CountriesTotal sample of countriesAfricaAsiaLatin America
(As share of total wages in the economy; in percent)
Central government(X)8.719.822.617.214.7
(s)(4.6)(9.9)(9.5)(11.9)(7.2)
(n)203514413
State and local government(X)11.63.61.76.2
(s)(6.7)(4.5)(2.0)(…)(5.4)
(n)1111515
General government(X)20.720.024.817.0
(s)(7.3)(7.3)(4.3)(…)(7.1)
(n)1111515
Nonfinancial public enterprises(X)4.68.48.69.7
(s)(4.0)(5.1)(6.6)(…)(5.1)
(n)510424
Public sector(X)22.932.0
(s)(7.9)(6.3)(…)(…)(…)
(n)56313
(As share of national income at market prices; in percent)
Central government(X)6.49.410.17.58.5
(s)(3.4)(4.3)(4.0)(4.7)(4.5)
(n)214318614
State and local government(X)8.21.50.72.7
(s)(4.7)(2.2)(0.9)(…)(2.9)
(n)1114616
General government(X)14.611.613.510.6
(s)(5.0)(3.3)(1.3)(…)(3.8)
(n)1114616
Nonfinancial public enterprises(X)3.73.84.14.0
(s)(2.9)(3.1)(4.5)(…)(2.3)
(n)612525
Public sector(X)17.016.718.515.0
(s)(6.2)(4.5)(5.7)(…)(2.6)
(n)58414
(As share of GDP: in percent)
Central government(X)5.27.98.36.07.2
(s)(2.6)(3.4)(3.3)(3.4)(3.4)
(n)215725718
State and local government(X)6.40.80.41.8
(s)(3.6)(1.6)(0.7)(…)(2.2)
(n)1122928
General government(X)11.58.89.59.1
(s)(3.7)(2.4)(2.4)(…)(2.5)
(n)1122928
Nonfinancial public enterprises(X)2.72.83.12.8
(s)(2.2)(2.2)(3.0)(…)(2.0)
(n)613526
Public sector(X)13.412.414.211.0
(s)(4.8)(3.8)3.6(…)(3.6)
(n)59415
Sources: Tables 25 and 26.

X = mean; s = standard deviation; n = number of observations in the sample.

Sources: Tables 25 and 26.

X = mean; s = standard deviation; n = number of observations in the sample.

Central government wages as a share of GDP tend to hover between 4 and 8 percent of GDP, with a higher share in the developing countries (7.9 percent) than in OECD countries (5.2 percent). One major contrast is in the share of state and local government wages, which averages 6.4 percent among the OECD countries and only 0.8 percent in the developing countries. This leads, not surprisingly, to a significantly higher share of general government wages in GDP among the OECD countries, averaging approximately 11.5 percent in contrast to only 8.8 percent in the developing world. Fewer data exist on wages in the nonfinancial public enterprise sector, but those available suggest the average share of wages in the OECD and non-oil developing countries to be comparable at 2.7 percent and 2.8 percent, respectively.

Another measure of the potential weight of government wage policy is the ratio of government wages to total wages in the economy. Among developing countries, central government wages are 19.8 percent of total wages, with the highest ratios in the African region and the lowest in Latin America (Table 4 and Appendix I, Table 26). Among the OECD members, central government wages are only 8.7 percent of the total; but if one also takes account of state and local government units, total general government wages reach 20.7 percent in the OECD, roughly comparable to the weight of general government wages among developing countries—20 percent.

It is of obvious interest to compare the share of government wages in total wages in the economy to the share of government employment in total nonagricultural employment. The comparison is a valid one only for economies where the “compensation of employees” in the national income accounts is derived primarily from nonagricultural sector employment.13 In making this comparison, important differences emerge between developed and developing countries.

In the OECD, the weights of general government wages and employment in total wages and total nonagricultural sector employment are similar (20.7 percent and 19.2 percent (see Table 1), respectively); in the developing countries, the employment share dominates the wage share (26 percent relative to 20 percent) suggesting that the average wage in the government sector of the developing countries is less than that in the nonagricultural private sector (including nonfinancial public enterprises). This finding suggests that government sector employees in those countries are not able to translate their strength in numbers into commensurate strength in their wage rates relative to that of their peers in the nonagricultural component of the private sector. (See also Section V.) Perhaps this reflects the fact that the government sector is used in many developing countries as a vehicle to absorb some of the unemployment, and the low productivity of underutilized government employees may be reflected in their lower relative wage rates. It may also reflect the view stated by the Malaysian Government, “Experience has shown that any increase in the pay of Government executives as a means to induce them to remain in the service will only be met by a corresponding or greater increase in the offers made by the Private Sector for the executives.”14

Finally, does the weight of public sector wages in national income affect the ultimate distribution of national income between labor and capital? In other words, is a high public sector wage share merely offset by a lower private sector wage share or does it bias the overall distribution of national income toward labor? A simple econometric test of this hypothesis has been constructed, as indicated in Table 5, suggesting that an increase in the share of central government wages in national income does lead to an almost concommitant increase in the total share of wages in national income. On a more limited sample, the share of state and local governmental wages in national income has no effect in the overall wage share. The higher the per capita income level, the higher the overall wage share.15

Table 5.Determinants of the Wage Share in National Income(t-statistics in parentheses)
Independent Variables Dependents VariablesShare of Central Government Wages in National IncomeShare of State and Local Government Wages in National IncomePer Capita Income (In thousands of U.S. dollars)ConstantR2(n)1
Share of total wages in national0.650.030.400.51
(1.62)(7.14)(9.01)(54)
Share of total wages in national income0.99−0.10.020.450.52
(1.68)(−0.17)(3.48)(6.43)(21)

n = number of observations in the sample.

n = number of observations in the sample.

Linkages Between Wage Expenditure and Total Public Expenditure

In an earlier study on public expenditure, the authors argued that the functional structure of public expenditure was a key determinant of the magnitude of public sector expenditure on any economic category of public expenditure, of which wages and salaries are among the most central.16 This is also a factor underlying the relatively higher importance of wages at the state and local government level. Whereas in many of the developed countries a significant share of central government expenditure relates to social cash transfer payments or is for services commissioned outside the public sector (via outside contracting), local government expenditure is devoted largely to the provision of services. The critical importance of the functional composition of expenditure emerges clearly in the sample of countries in this study.

For example, if one relates the share in GDP of central government wages to the share in GDP of total central government expenditure, one observes (Table 6) a clear positive relationship up to a per capita income of US$1,000 and then a sharp negative relationship at higher per capita income levels. The significance of this reversal is the increasingly important role played by government subsidies and transfers as per capita income rises. If the central government’s wage share in GDP is related to the share in GDP of public expenditure exclusive of subsidies, there is a uniformly direct relationship, regardless of the per capita income level. Relating the wage share in GDP to the share in GDP of central government expenditure on different functional categories of expenditure also supports this hypothesis. Expenditure on education, public administration, and defense prove to be wage intensive; expenditure on social security, health, and economic services prove to have little impact on the wage share. (See Table 6.)

Table 6.Functional Expenditure Determinants of Central Government Wage Expenditure1(t-statistics in parentheses)
Expenditure on
Independent Variables /Dependent VariablesTotal Central Government ExpenditureTotal Central Government Expenditure (Excluding Subsidies)Economic servicesEducationHealthSocial security welfareDefensePublic administrationConstantR2(n)2
Central government wages:
For countries with per capita income ≤ US$1,0000.143.930.30
(5.43)(5.08)(71)
For countries with per capita income > US$1,000−0.473.930.30
(−2.95)(5.08)(71)
Central government wages0.252.360.49
(7.46)(3.42)(65)
Central government wages−0.030.77−0.17−0.070.140.343.300.46
(−0.15)(3.30)(−0.68)(−1.15)(1.67)1.64(3.35)(55)

All variables are taken as a share of GDP.

n = number of observations in the sample.

All variables are taken as a share of GDP.

n = number of observations in the sample.

    Other Resources Citing This Publication