Wage policies in the public sector during wholesale privatization

Why should pay differ between the public and private sectors?

Policymakers need to ensure that wage determination in the public sector provides competitive compensation relative to the private sector. If the public sector lacks the capacity to match private sector wages, it will have difficulty hiring and retaining a high-quality workforce. Consequently, the quality of public services may deteriorate, and problems such as corruption and worker turnover could become more prevalent, adversely affecting the entire economy. However, if the public sector pays higher wages, the private sector could face difficulties in recruitment and encounter upward pressure on wages as workers queue to enter the public sector. Further adverse effects on employment could be caused by increases in taxes or reductions in public investments necessary for economic growth in order to finance high public wages. A rise in public sector wages during political cycles or economic expansion may be hard to reverse in periods of recession and may lead to an increase in informal sector activities.

Simple comparisons of raw public sector wages with those in the private sector, as are typically reported in policy documents, can be very deceptive. Empirical evidence shows that higher wages in public jobs can largely be explained by job composition and worker characteristics. Public sector establishments are, on average, larger than private sector firms, and the majority of jobs in the public sector in OECD countries are also now clerical, with only a very small fraction of manual jobs. Public sector workers are also found to be, on average, better educated, older, and more risk averse than private sector workers. When these characteristics are taken into account, the public sector wage premium either decreases or turns into a penalty relative to the private sector.

A variety of methodological approaches are used to control for compositional differences between the two sectors. One common approach is to pool data across workers in both sectors in the wage equation that includes a set of worker and job characteristics, and a dummy variable indicating whether a worker has public sector status or not. A second, less restrictive, approach estimates wage equations for the public and private sector samples separately. In this case the wage gap is decomposed into two components. One is the difference in worker and job characteristics; the other is the difference in the returns to characteristics, and represents the wage gap upon which policymakers should focus. In a third approach, the wage gap estimates can be obtained using methods that correct for potential non-random sector selection based on unobserved worker characteristics (such as ability, risk aversion, and family influence). These approaches can be applied at the mean and at different points of the wage distribution.

Although most of the characteristics that explain compositional disparities between sectors are common to both transition and developed market economies, there are some important differences. For example, differences in the union status between the public and private sectors are important in market economies, whereas unions in transition economies are found to have only a modest influence on wage-setting [4].

An important feature in transition economies is a potential large disparity in non-wage benefits between the two sectors, such as meal and travel allowances, subsidies, payments in-kind, and bonuses, which must be accounted for when analyzing sectoral wage gaps [4]. Moreover, wage arrears are widespread in economic transition and may also affect the wage gap due to non-random distribution between sectors.

Systematic wage differences between public and private sectors could potentially exist due to intrinsic differences in wage-setting behavior. Specifically, public sector workers may be underpaid relative to their private sector counterparts if the public sector holds market power in wage-setting as a result of its being the dominant employer. In market economies, this is typical for selected sectors, such as education, health care, law enforcement, and public administration. In communist economies, the public sector was the employer of nearly the entire workforce. Prior to transition, wages were not directly linked to enterprise performance, but were centrally determined with minimum and maximum wages, which created a very egalitarian wage distribution. Economic transition meant that although the government might initially continue to impose restrictions on pay, workers could opt to change their sector of employment, as there would now be a private sector alternative in which wages would be determined based on profit maximization. The literature suggests that private sector expansion and an increase in demand for skills during transition caused significant growth in wage inequality (faster than in other OECD economies) [3]. Wage differentials were one of the key variables in labor markets in transition economies.

Changes in the public sector wage gap during transition

A number of studies carried out at the initial stages of economic transition found that public sector workers earned, on average, lower wages than their private sector counterparts. This result is robust across different methodological approaches, definitions of public sector, wage definitions (monthly, hourly, before or after tax), gender, and data sources (whether self-reported or employer-provided). In most countries the negative gap between public and private sector wages narrows over time [5].

The Illustration presents results from studies that estimate the mean public–private sector wage gap, conditional upon observed worker and job characteristics. The selection of studies for this comparison was based on the use of a common method of estimation across countries in two transition periods: the 1990s and early 2000s. The periods relate to ownership transformation and public sector wage reforms respectively. In all sources used for the Illustration, regression specifications include experience and/or tenure and education, as well as a range of other relevant variables. However, the purpose of the Illustration is just to show the change in the public–private wage gap between the two periods for each of the selected economies, rather than to make a cross-country comparison of the size of the wage gap. This is primarily due to available data sets that determine sample variables and sector definitions used in the estimations. For example, in studies for Hungary and Ukraine the public sector includes budgetary institutions (public administration, health, and education), whereas for Poland, Serbia, and Croatia, it includes both budgetary institutions and state-owned enterprises.

Although the Illustration presents estimates from self-contained studies that focus on different countries and use different sources of microdata, a common story can be drawn: there was a negative public–private sector wage differential during the 1990s that declined in the early 2000s. The study on Poland finds that the gap changed from -11% to 0% [3]. In Serbia, it changed from -18% to -9% [4], in Croatia from -4% to +6%, and in Hungary from -20% to -16% [1]. The study on Ukraine estimates a change from -21% to -16.5%, and shows that estimates obtained using longitudinal data remained very close to cross-sectional estimates (-20.5% from 1997–2003) [2]. The same study (not shown in the Illustration) finds that the difference was greater for female employees, indicating that sorting into the public sector based on individual preferences is greater for female than for male employees.

Other studies that correct for sector selection also estimate negative public–private wage gaps. A further study on Poland reports the wage gap to be -7% for men and -10% for women in 1996, after correcting for lower entry costs into the private sector for younger individuals [6]. Research on Yugoslavia estimates a -9.4% wage gap for male employees in 2000 after correcting for differences in sectoral job security and associated benefits [7].

Why is it important to look beyond the comparisons at the mean?

A closer look at data from the same studies used in the Illustration suggests discrepancies in the size of the public wage gap at three different points of the wage distribution.

Figure 1 reveals an increasing public sector wage penalty along the wage distribution. In other words, workers with high wages are more underpaid than those with low wages when both are compared with their private sector counterparts. Whilst public sector wage compression is also typical for developed countries, a comprehensive literature survey suggests greater compression in transition economies [5]. Consequently, higher paid workers had greater incentives than lower paid workers to search for better opportunities in the private sector. The implication of this for policymakers is clear: appropriate wage reforms are necessary to retain the most skilled workers, as well as to prevent corruption and deterioration in the quality of public services.

Effects of changes in public sector wage policies

A number of studies have analyzed changes in public sector wage-setting in transition economies (see Wage reforms in Serbia and Hungary). One such study on Serbia finds that the public–private wage gap at the mean in 2004–2008 was close to zero for workers with tertiary education, while a premium emerged for workers only educated to either a primary or secondary level (14% and 13% for males respectively) [4]. Similar research for Hungary shows that the wage gap at the mean changed from a penalty in 2000 into a premium in 2004 for workers educated to below tertiary level (-13% turns into 14% for workers with primary and vocational education, and -22% turns into 10% for high school graduates), while those with tertiary education saw the largest improvements relative to their private sector counterparts (the gap changed from -43% in 2000 to -4% in 2004) [8]. This suggests an increase in public sector wage inequality.

The literature from Estonia, Serbia, and Russia consistently reports faster growth of public sector wages relative to private sector wages for workers at the top end of the wage distribution than for workers at the bottom end [4], [9], [10]. For example, research on Estonia estimates that in 1989 a -23% public wage gap existed at the 10th percentile, and a -77% gap at the 90th percentile. Ten years later, in 1998, workers at the 10th percentile enjoyed a 13% public sector premium, whereas those at the 90th percentile saw the gap reduce to -9.4%. In Russia, policy changes to redistribute wage funds towards the most skilled government workers were undertaken only recently, in 2012. Nevertheless, recent research has also shown that the public sector wage penalty has decreased, and that the largest relative gain is observed for those placed in higher percentiles [10].

More detailed analyses by educational qualification and gender in Hungary reveal that university graduates in the 10th percentile saw the greatest relative financial gain between the 1990s and early 2000s when compared with all other groups of workers [1]. Wage reforms are found to have caused a premium at the 10th percentile, while the penalty at the 90th percentile changed little, as shown in Figure 2 [12]. For policymakers, the lesson from this evidence is that although Hungarian public sector wage reforms increased inequality between the average wages of workers with different educational qualifications, revisions of wage scales led to wage compression among workers with the same educational qualification, especially for university graduates.

Similar analyses for Serbia find that there was an increase in relative public sector wage dispersion [11]. As shown in Figure 2, in 2004–2008, Serbian public sector workers with primary education mostly collected a uniform premium along the wage distribution. Among university graduates, the public–private sector wage gap was eliminated for workers at and below the median, but this was not the case for workers with wages above the median wage. This suggests that the public sector wage distribution for university graduates widened only between median and lower wages.

In sum, several adverse effects should be highlighted based on the evidence presented in Figure 2. A large starting wage premium for high-skilled workers in the 10th percentile is likely to attract more competent young university graduates to enter the public sector, as in Hungary, while the private sector will face hiring difficulties as a result. Additionally, increasing penalties at higher percentiles of the wage distribution is likely to cause greater worker turnover, increase disincentives, foster moonlighting, and lead to underperformance at the more senior level. However, higher public sector than private sector wages for workers educated to less than a tertiary education, as in Serbia, will likely reduce public sector efficiency, strain its finances, and hinder private sector job creation.

The 2008 global financial crisis further challenged public wage bill sustainability and revealed weaknesses of pre-crisis wage reform processes in transition countries. An analysis of EU countries reported that the average for the public–private wage gap was 3.6% in 2010 [13]. The same study found that Poland and Slovenia were the only EU post-transition countries to have a positive wage gap (6.5% and 5.4% respectively). A negative public–private wage gap was still mainly concentrated among countries that went through transition. Specifically, the gap was -16% in Hungary and -4.8% in the Czech Republic; all other transition countries fell somewhere in between, except for Romania and Latvia, where there was no gap [13].