Journal Issue

Romania: Selected Issues

International Monetary Fund
Published Date:
June 2007
  • ShareShare
Show Summary Details

II. Wage Dynamics in the Romanian Economy26

A. Introduction

58. Wage restraint has been an important element of Romania’s stabilization process (Figure 1). By keeping a lid on public-sector wage increases—through prudent wage increases in the general government and strict control of the wage bill of the state-owned enterprises (SOEs)—and approving modest increases in the minimum wage, the growth in the economy-wide real wage has been moderate. This has contributed to modest increases in unit labor costs which have been below productivity gains for most of Romania’s recent history. As a result, Romania has achieved substantial competitiveness gains and export performance has been strong, despite a sharp appreciation of the exchange rate. In a related vein, a prudent incomes policy—in association with tight fiscal (until late-2006) and monetary policies—has contributed to a sharp decline of inflation.

Figure 1.Romania: Wage and Fiscal Developments, and Macroeconomic Performance, 2000-06

1/ Core 2 CPI excludes administered prices, volatile food items, and fuel.

Source: Romanian authorities and staff estimates.

59. However, the recent loosening of public-sector wage policy has raised concerns about Romania’s ability to achieve sustainable strong growth in a low-inflation environment. Since end-2004, wage increases in the government sector have considerably outpaced ones in the private sector. At the same time, public-sector employment has been expanding (partly on account of the need for EU-accession related hirings), in a labor market characterized by low unemploymnt, participation, and employment rates. Moreover, the monitoring of the SOEs’ wage policy eased this year. These developments raise concerns about whether private-sector wages—which have so far increased only modestly—will follow these trends and exhibit significant increases. These concerns are exacerbated as the Romanian economy has shown persistent signs of overheating.

60. A number of other institutional factors of the wage-determination process are also cause for concern. Although Romania has achieved low single-digit inflation, multiple increases in government wages continue to take place every year. Moreover, due to the timing of government-wage agreements, most wage increases are not included in the budget approved by parliament, thus creating the need for additional budget allocations and budget revisions during the course of the year. As a result, in many instances, additional wage allocations displace other government spending, including, most importantly, capital. Finally, although government employment has been expanding steadily, there is no well-specified medium-term plan for the reform of public administration.

61. Evidence from other countries points to a generally strong impact of publicsector wage policy on private-sector settlements. Demekas and Kontolemis (1999) develop a two-sector labor market model and by applying it to Greece—a country with a relatively large public sector—show that increases in government wages lead through worker-flow dynamics to increases in private-sector wages and, therefore, directly to higher unemployment. Mizala and Romaguera (1995) find that in the case of Chile, private-sector wages have followed those of the public sector in a time path, and that the leadership role of public wages has gradually become weaker as state intervention in the labor market declined. In a system of joint wage determination of private, public, and local sector wages in Sweden, Jacobson and Ohlsson (1994) found that private sector wages lead public-sector settlements. Finally, Bemmels and Zaidi (1990) looked into the wage leadership hypothesis in the Canadian industry and were able to identify that certain institutional forces (e.g., unionization or efficiency wages) of some sectors contributed to those sectors being leaders of the wage determination process.

62. The objective of this paper is to examine recent wage developments in Romania, and to test the links between the wage determination process of the public and private sectors. Section B presents a number of stylized facts regarding recent wage and employment developments, and the wage-setting process. Section C presents the results of econometric analysis aimed at testing the links between wages in various sectors of the economy; and Section D summarizes the results and provides policy recommendations.

B. Stylized Facts

Wage and employment developments

63. Wage moderation has been an important element of Romania’s stabilization process, but recent developments raise concerns. Up until mid-2006, the gross economy-wide real wage increased only moderately. This was a reflection of continued enterprise restructuring and took place despite relatively tight labor market conditions, high migration to western Europe, and expanding public-sector employment. Romanian (net) incomes benefited from the cut in the personal income tax to a flat rate of 16 percent on January 1, 2005, while the currency appreciation led to higher purchasing power and wealth effects. However, since mid-2006, the economy-wide real wage has been increasing fast, reflecting continued tightening of labor market conditions and a pronnounced lack of educated and skilled workers, but also driven by sharp increases in government wages. This development is taking place at a time when the economy is overheating—as evidenced by the widening external imbalance—while inflationary pressures are building. Non-government credit growth is very strong; and fiscal policy is contributing to demand pressures following the loosening that started at end-2006 and in light of the envisaged relaxation this year. At the same time, the gap between real wage increases and productivity gains has been closing fast, thus raising concerns about enterprise profitability and Romania’s competitiveness edge.

64. Substantial increases have been provided to the government sector. Following a trend that started in late-2004, the government has been giving substantial wage increases in all sectors of the general government, namely, public administration, education, health, contractuals, the police, and the army (Table 1). Government officials have justified the magnitude of such increases as necessary to attract high-quality staff in the public sector in line with Romania’s objective to improve the efficiency in the provision of government services. They also cite the high competition for jobs with the private sector, which has been expanding rapidly. At the same time though, the loose government-wage policy has been greatly influenced by strong labor unions (particularly in the education sector), in the context of a tense political environment.

Table 1.Romania: Wage Policy of the General Government in 2004-07(In percent)
Public AdministrationContractual-Civil AdministrationArmy
SalaryCumulative IncreaseSalaryCumulative IncreaseSalaryCumulative Increase
Increase(In each year)(In 2004–07)Increase(In each year)(In 2004–07)Increase(In each year)(In 2004–07)
SalaryCumulative IncreaseSalaryCumulative IncreaseSalaryCumulative Increase
Increase(In each year)(In 2004–07)Increase(In each year)(In 2004–07)Increase(In each year)(In 2004–07)
Sources: Romanian authorities and staff calculations.
Sources: Romanian authorities and staff calculations.

65. As a result, the government’s wage bill has increased substantially. From 4.8 percent of GDP in 2004, the general government payroll increased to 5.4 percent and 6.1 percent in 2005 and 2006, respectively. Generous wages increases, increases in employment, and various bonuses contributed to such outcomes. In 2006, average statutory wages increased by 9–15 percent and employment expanded by 3⅓ percent; however, on account of increases in bonuses and payment of retroactive wages for teachers and judges due to legal settlements, the wage bill increased by 36 percent in nominal terms.

65. Government wages are on average higher than private-sector ones (Table 2). Having benefited from multiple sharp increases, the average wages in public administration and education are now 72 and 16 percent higher than the economy-wide average wage. Regarding public administration, in many countries it is the case that wages are higher than the rest of the economy, due in part to the fact that many public administration employees have usually completed tertiary education. However, the rate of wage increases is not at all usual.

Table 2.Romania: Gross Wage by Sector(In percent of the economy-wide average wage)
AgricultureIndustryConstructionTradeFinancial IntermediationPublic AdministrationEducationHealth
Sources: Romanian authorities and staff calculations.
Sources: Romanian authorities and staff calculations.

66. These wage developments have been taking place in a tight labor market. Romania’s unemployment rate is among the lowest in the European Union (Figure 2), while the labor market is particularly tight for skilled and educated workers. Labor market tensions are also exacerbated due to the very low labor-market participation and employment rates. Despite some increase in recent years, both rates (61 and 56 percent, respectively in 2005) are among the lowest in the European Union (Figures 3 and 4). A recent study by the Ministry of Labor (2006) looked into the dynamics of the labor market and noted the tensions stemming from the widening gap between labor demand and supply in particular segments of the labor market.

Figure 2.Romania: Unemployment Rates in Selected EU New Member States, 1991–2005

(In percent)

Source: AMECO.

Figure 3.Romania: Labor Force Participation in Selected EU New Member States, 1991–2005

(in percent of working-age population)

Source: AMECO.

Figure 4.Romania: Employment in Selected EU New Member States, 1991–2005

(in percent of working-age population)

Source: AMECO.

67. Public-administration employment policy has been crowding out private sector employment. While total economy-wide employment has been increasing at a slow pace (and industry employment has been declining in line with the ongoing enterprise restructuring), employment in the government sector has been growing at a much-faster rate. This outcome is driven by the very fast increase in employment in public administration. For example, in 2006, although the authorities had committed to freezing vacant positions (except for EU-related hirings) to provide resources for wage bill overruns, government employment increased by almost 30,000 (3⅓ percent). Although many of those positions may reflect hirings associated with commitments to the EU (e.g., to improve border protection and the capacity for absorbing EU funds), such an expansion of government employment is taking place without a well-specified plan for the reform of public administration. Thus, considerable concerns exist about the efficiency in the provision of services by public administration.

68. The gap between Romanian and average EU wages has narrowed recently. Wage increases of recent years (which exceeded average EU wage increases) coupled with significant nominal appreciation of the leu have led to a narrowing of the gap of Romanian wages (measured in euros) from the EU-average wage. During 2000-06, the average compensation of Romanian employees exhibited the largest increase among EU countries and almost doubled. Although it is among the lowest, it is now comparable with a number of recent new member states of the European Union.

Source: AMECO database.

69. The minimum wage remains low. Owing mainly to prudent increases in the minimum wage during 2004–06 (nominal increases of 12, 11, and 6 percent, respectively), the minimum wage had little impact on the behavior of the economy-wide wage. This is the case notwithstanding the fact that a relatively high (12 percent in 2004; and among the highest in the EU) share of full-time employees have earnings on the minimum wage (Eurostat (2006)). However, the recently-approved significant increase in the 2007 minimum wage of 18 percent is going to increase the minimum wage relatively to the average economy-wide significantly, thus possibly providing a signal for high privatesector wage increases.

Source: Eurostat.

The government wage-setting process

70. There is no unified pay system in the Romanian public administration, as the legal framework differentiates between several employment categories. As World Bank (2006) indicates, in addition to the general principles stipulated by the Labor Code and other normative acts, various regulations (both general and specific) influence substantially pay structures and employment policy. This structure complicates and segments the wage bargaining process and salary setting, resulting in large discrepancies across and within public sector categories.

71. The timing and duration of wage negotiations presents challenges for the budget process and fiscal management. Negotiations for the wage increases to be effected the following year usually begin toward the end of the year and often get protracted and last well within the new year. As a result, the budget that is approved by Parliament does not include fully the impact of agreed wage increases. Despite attempts by the Ministry of Finance to create some space in the wage bill for wage increases, it is usually the case that budget rectifications during the year have to provide additional resources for payroll. For example, in 2006 the original budget approved by parliament in December 2005 included an allocation of 5.5 percent of GDP for wages. This allocation was successively increased to 5.6, 5.8, 5.9, and 6.0 percent of GDP during each of the four budget rectifications, while the outturn for the year reached 6.1 percent. Another remnant feature of Romania’s long tradition of high inflation is the existence of multiple wage increases. Perhaps as a way to deal with high inflation and preserve purchasing power, the government has traditionally offered more than one wage increase during the year to its employees. However, this “tradition” continues even now, although Romania has achieved low single-digit inflation.

72. The base salary—which varies widely across public institutions—is only one component of the total compensation package. Due to the complicated legislation and various regulations, the pay structure of the government varies considerably across government institutions (Table 3). The most important element is the limited role and the significant variation of the base salary (ranging from 52 to 67 percent of total compensation). The rest of the compensation is accounted for by various bonuses, which are provided on the basis of both objective and subjective criteria. This compensation structure and the magnitude of bonuses in the total compensation package raise questions about the effective budgeting, and the transparency and monitoring of the wage bill, as well as the objectivity in providing such rewards.

Table 3.Romania: Pay Structures for Civil Servants (by type of public institution)(In percent of total compensation)
Pay ElementType 1Type 2Type 3Type 4
Base Salary60525267
Managerial bonus4322
Performance salary2222
Seniority allowance13111013
“Dispozitiv” allowance0200
Loyalty allowance8000
Various other bonuses8549
Note: Type 1 represents public institutions subordinated to the central Government; type 2 covers ministries; type 3 covers public institutions subordinated to ministries; type 4 covers local public institutions.Source: World Bank (2006) on the basis of the Hay Group report.
Note: Type 1 represents public institutions subordinated to the central Government; type 2 covers ministries; type 3 covers public institutions subordinated to ministries; type 4 covers local public institutions.Source: World Bank (2006) on the basis of the Hay Group report.

Wage linkages

73. The incomes policy of the public sector could affect economy-wide wage settlements in a number of ways. First, since agreements on increases in the government wage, the minimum wage, and the wages of SOEs usually take place before similar settlements take place for the private sector, sharp increases in public-sector wages could provide a strong signal (demonstration effect) to the private sector. Second, in countries where the government employs a large and growing number of staff, significant increases in government wages, lead through worker-flow dynamics to increases in private-sector wages. This mechanism is particularly strong in countries with very low unemployment. In a related vein, competition by the government for the hiring of educated and skilled employees could create tensions for particular segments of the labor market. Finally, since Romania has had a long history of high inflation, wage indexation mechanisms have been quite prevalent, were based on historical inflation, and, as noted earlier, came in multiple rounds.

C. Econometric Analysis

74. A bivariate vector autoregressive (VAR) system is estimated in order to model sectoral wage behavior. Such a system is a popular framework for estimating and evaluating economy-wide models, since, among other things, it provides a flexible forecasting model, forms the basis of Granger (1969) causality testing, and can be extended to incorporate various restrictions in the form of structural VARs. The key characteristic of the model is that it expresses the wage of a sector of the economy as a function of: (i) its own lags and (ii) lags of wages in other sectors. Thus, in the case of private and government wages, the bivariate VAR takes the form:

where: WPRIVt and WGOVt are the monthly rate of growth in the real private- and the government-sector wage, respectively; the explanatory variables are lags of the two wages, and u1t and u2t are the disturbance terms. A VAR can be estimated by applying ordinary-least squares to each equation, which produces efficient parameter estimates. Although economic theory could provide some a priori judgment about the number of lags, the optimal number of lags (r and s) in the VAR could also be chosen by employing three information criteria: Akaike, Schwarz, and Hannan-Quinn.27

75. Having estimated a VAR for wages, Granger-causality testing could provide information about the relationship between movements in wages of different sectors. In the Granger sense, a time series Xt is said to cause another series Yt if the future values of Yt can be better predicted when the information contained in the series Xt is included rather than excluded. Granger-causality tests are based on testing the joint significance of the lags of each variable in the system, apart from its own lags. Thus, a test that an increase in the government wage WGOV Granger-causes an increase in the private-sector wage WPRIV is given by testing the hypotheses:

H0 : γ1 = γ2 = …= γs = 0 (No Granger causality)

H1 : at least one of the restrictions fails,

where the parameters γ1, γ2, …γs are defined in the first equation of the VAR (1). The reverse test that an increase in the private-sector wage WPRIV Granger-causes an increase in the government wage WGOV is given by testing the hypotheses:

H0 : δ1 = δ2 = …= δr = 0 (No Granger causality)

H1 : at least one of the restrictions fails,

where the parameters δ1, δ2, … δr are defined in the second equation of the VAR (1).

76. We employ this methodology to test whether wage increases in one sector of the economy lead others. By estimating bivariate VARs we examine whether a Grangercausality relationship exists among wages in government, state-owned enterprises, and the private sector. Specifically, we examine two hypothesis: (i) whether government wage developments affect private-sector settlements and visa versa; and (ii) whether developments in state-owned enterprises wages affect private-sector and government wage settlements and visa versa. We use monthly data on gross wages (deflated by consumer prices) compiled by the Romanian Statistical Office for various sectors of the economy. We proxy the government wage as a weighted average of wages in public administration, education, and health by using employment as weights. Data span the 1993–2007 period, with the exception of wages in the SOEs, which begin in 1997. Since wage dynamics have evolved over time as part of Romania’s transition to a market economy, we test the developed hypotheses for various time periods. Seasonal dummies were included in models where they were found to be significant (e.g., a December dummy was important in explaining end-year bonuses provided by the government sector and SOEs).

77. The results of causality tests indicate strong links between the wage-setting behavior of the government and the private sector (Table 4). By employing the entire sample (which starts from the beginning of Romania’s transition) we find that private sector wages lead wages in the government sector. Specifically, the null hypothesis that the coefficients of the real wage increase in the private sector are zero is rejected in almost all model specifications. However, in the reverse case, for samples beginning until 1997, we were not able to reject the null hypothesis that the coefficients of the real increase in the government wage are zero. This result points to the conclusion that government wage increases do not cause (in the Granger sense) private-sector wage increases. However, for all samples covering the 1998–2006 period, bi-directional causality is obtained. In other words, in recent years and as Romania has been moving along its transition to a market economy, and labor market conditions have become tighter, the wage policy in the government sector appears to influence wage settlements in the private sector.

Table 4.Romania: Granger-causality Tests for Private Sector and General Government
Causality directionChi-square testNumber of lagsProbability
Private Sector→Government4.3910.036
Government→Private Sector0.0110.916
Private Sector→Government6.2120.045
Government→Private Sector0.5920.743
Private Sector→Government7.4420.024
Government→Private Sector0.9320.630
Private Sector→Government10.9450.050
Government→Private Sector3.0250.670
Private Sector→Government19.12120.086
Government→Private Sector18.59120.099
Private Sector→Government23.62140.051
Government→Private Sector61.30140.000
Private Sector→Government1.0720.586
Government→Private Sector7.4720.024
Private Sector→Government32.05140.004
Government→Private Sector45.45140.000
Private Sector→Government29.16150.015
Government→Private Sector45.58150.000

78. Wage policy in the state-owned enterprise sector does affect private-sector wage developments (Table 5). Granger tests indicate causality from SOEs’ wage increases to wage increases of the private sector. Specifically, the null hypothesis that the coefficients of the real SOEs wage increases is zero is rejected for all model specifications; these coefficients are statistically different from zero at a significance level of less than 8 percent. However, that is not case when reversing the equation with the coefficients of the lags of the growth rate of private-sector wages, thus implying that private-sector settlements had no influence on wage settlements in the SOEs sector. Finally, a test of the causality between government and SOEs’ wages points to a strong leadership role of government wage policy on wage developments in the SOEs, but not visa versa. One interpretation has to do with the timing of the wage increases, since government wage policy is usually determined before the one of the SOEs.

Table 5.Romania: Granger-causality tests for Private Sector, Public Enterprises, and Government
Causality directionChi-square testNumber of lagsProbability
Private Sector→Public Enterprises5.3630.147
Public Enterprises→Private Sector22.6430.000
Private Sector→Public Enterprises5.79120.926
Public Enterprises→Private Sector20.28120.062
Private Sector→Public Enterprises23.52180.171
Public Enterprises→Private Sector27.24180.075
Government→Public Enterprises16.3930.001
Public Enterprises→Government3.7130.294

D. Conclusions

79. This paper examined the dynamics of wage determination in Romania, with particular focus on the influence of government-wage policies on private-sector settlements. The analysis showed that private- and public-wage settlements are closely linked, in the sense that private sector wage increases affect public wage increases and visa versa. This implies that although prima facie, the private and public sector determination processes are separate, they are effectively linked. Wage agreements in the government provide a signal to private-sector wages, on account of their timing, and the fact that the government competes with the private sector for educated and skilled staff. It is the interaction in the competition for jobs, particularly in an environment of low unemployment, labor participation and employment rates, that provides additional impetus to the interaction between government- and private-sector wages. The paper also showed a strong leadership role of the SOEs’ wage policy on private-sector wage agreements.

80. The results of the paper are particularly relevant for Romania’s current position on the business cycle. In the context of an economy that is overheating—as evidenced by a widening current account deficit, very strong demand pressures, and building of inflationary pressures—the government’s wage policy has an important role to play. This role becomes even more crucial as public-sector employment has been rising under very tight labor market conditions. Thus, a prudent wage policy in the broad public sector (general government and SOEs) could provide an appropriate signal to wage settlements in the private sector and weigh heavily on the wage determination process.

81. These results are also important as Romania improves its institutional setting of macroeconomic policy. With the transition to a low-inflation environment, it is important that the wage-setting process changes in various ways. For example, public sector wages should increase only once a year, and the agreed wage increase should be moderate and take into account expected inflation. Moreover, government wage increases need to be approved before the budget is approved by parliament, and, if possible, be integrated into a credible medium-term fiscal plan.

82. As part of convergence, Romanian wages will rise, but policymakers need to watch carefully the speed of convergence. Although consecutive sharp wage increases coupled with leu appreciation have led to a narrowing of the wage gap from EU average wages, Romanian wages (in euro terms) are still low. There is no doubt that, as part of income convergence, economy-wide wages will move toward the EU average. The speed of convergence, though, could have important macroeconomic implications (including supporting sustainable strong growth in a low-inflation environment). Thus, due attention is needed to ensure that rapid wage increases do not outpace productivity gains, and in turn hurt Romania’s competitiveness position.

83. Finally, a well-specified reform plan is needed to strengthen the quality and efficiency of public administration. Maximizing the benefits of membership in the European Union requires establishing a modern public administration that responds to the country’s priorities, including the need for high absorption of EU funds. Better alignment of functions within public administration, an upgrade of skills of public-sector employees, and efficient and monitorable rules for the hiring, promotion and dismissal of staff are essential elements of a reform plan. Moreover, wages need to better reflect skill needs, otherwise a ballooning wage bill will only undermine the aim of establishing a modern public administration and improving the efficiency in the provision of public services.


Prepared by Costas Christou.

The Akaike (1969) criterion in general chooses long lag structures, whereas the Schwarz one tends to choose very short lag structures. In cases of different suggested optimal lags, researchers usually use the Hannan-Quinn criterion as an in-between measure. In the models estimated in this paper, we follow the same approach, but also report results using all criteria in case of differences in the derived results.

Other Resources Citing This Publication