Greece: Selected Issues and Statistical Appendix

This paper provides model-based projections of inflation, and quantifies the impact of the factors that determine inflation. The reasons behind the poor labor market performance in Greece and the remedial policies are discussed. The new economy is at a relatively early stage of development in Greece compared with most industrial countries, but growing rapidly. The key issues facing the Greek banking sector, against the background of its recent performance and a rapidly evolving market and regulatory environment, is discussed. The statistical data are also presented.

Abstract

This paper provides model-based projections of inflation, and quantifies the impact of the factors that determine inflation. The reasons behind the poor labor market performance in Greece and the remedial policies are discussed. The new economy is at a relatively early stage of development in Greece compared with most industrial countries, but growing rapidly. The key issues facing the Greek banking sector, against the background of its recent performance and a rapidly evolving market and regulatory environment, is discussed. The statistical data are also presented.

III. Greece’S Labor Market—Graippling with High Unemployment25

A. Introduction

69. With Greece’s successful accession to the euro area, attention has shifted to achieving a real convergence of living standards in Greece with those in its partner economies. In this respect, and in comparison with other lesser advanced peripheral euro-area economies, Greece has made only limited progress. One reason for this has been the inadequate contribution from labor in the growth process. While the unemployment rate has fallen for the last few years in the euro area, it has been rising steadily for the last half decade in Greece, and now exceeds the former by a wide margin. The focus of this paper is to explore the reasons behind the poor labor market performance in Greece, and to discuss briefly remedial policies.

70. This chapter’s main findings are as follows. While over the last two decades Greece has generated employment at a rate in excess of the euro-area average (although not so for the second half of the 1990s), even faster labor force growth in Greece has resulted in a substantial increase in the unemployment rate, in contrast to declining rates in many partner countries. Developments in Greece reflected an interaction between “shocks” to the labor market and Greece’s particular institutional setting (see Blanchard, 1998). The shocks included declining agricultural sector employment, rising female participation rates, and immigration, as well as industrial restructuring and a halt to rapid growth in public employment. Greece’s labor market structures and institutions are comparatively rigid, with high relative minimum wages, strict employment protection legislation, and shortcomings in the educational system. These interactions have resulted in increasing unemployment with lengthening duration, especially among women and the young. The authorities have responded to these developments with some reforms that have been in general useful, but fail to address some of the key structural weaknesses in the labor market. The paper concludes with a review of additional reform options.

71. The chapter is organized as follows. Section B examines Greek labor market developments from an international perspective, followed by a review of Greek labor market institutions (Section C), identifying similarities and differences with other, mainly European, economies. Section D discusses the interaction of economic shocks and institutions in accounting for past labor market developments, making use of measures for determining structural labor supply and demand. Section E reviews the Greek authorities’ policy responses to date, and concludes by discussing additional measures that may be needed to achieve a major reduction in unemployment, and to facilitate the convergence of real living standards.

B. Labor Market Developments

72. Greece’s current economic expansion has entered its eighth year, with rates of growth exceeding those in the euro area for the last five (Figure 1).26 But despite this generally favorable development, employment growth in Greece in the current expansion has lagged that in the euro area.27 More troubling, while the euro-area unemployment rate fell for the last three years, to below 9 percent by end-2000, Greece’s unemployment rate rose steadily throughout the previous decade, to 11.9 percent in 1999 (data for 2000 are not available).28

Figure 1.
Figure 1.

Per Capita GDP at Current Prices

(Purchasing Power Standard, EU 15 = 100)

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Source: European Commission (2000a).

73. When viewed from a longer perspective, employment in Greece over the last two decades has risen by more than in the euro area on average, and has recorded one of the highest growth rates among EU economies (Figure 2).29 But the expansion of the labor force over this period has been even larger, resulting in an increasing unemployment rate. The labor force’s growth itself reflected a number of factors. While the growth rate of the population, and of the domestic working-age population in particular has been falling, it was temporarily supplemented by sizable net immigration in the late 1980s and early 1990s, including by ethnic Greeks from the Black Sea basin (Figure 3). With the waning of this factor, however, the growth of the working-age population approached zero by the end of the 1990s. The second major factor influencing labor market growth was a slowly rising labor participation rate, with an increasing female rate more than offsetting a previously declining, and now largely constant, male rate (Figure 4). However, as in other euro-Mediterranean economies, the female participation rate remains well below the euro-area average.30 The youth (aged 15–24 years) participation rate in 1999, at 36.4 percent, is also well below the EU average of 46.6 percent, although a review of enrollment rates suggests that this pattern is not explained by an unusually large fraction of the cohort participating in education.

Figure 2.
Figure 2.

Labor Market Developments, International Comparisons

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: National Statistical Service of Greece; OECD Analytical Database; and Fund staff calculations.
Figure 3.
Figure 3.

Greece: Population Growth Rates

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Source: OECD (1999b).
Figure 4.
Figure 4.

Greece: Employment, Nonparticipation and Unemployment Rates

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: National Statistical Service of Greece; and Fund staff calculations.

74. A number of notable changes in employment patterns have emerged in the past two decades. Agricultural sector employment has been steadily declining over the period, with the drop averaging about 1 percent of total employment a year over the last three years; nevertheless, the sector still accounted for 17 percent of total employment in 1999 (Figure 5). Industry’s relative employment share remained constant at about 20 percent until the early 1990s, when long-delayed restructuring occurred, with a resulting decline in the share to about 16 percent in 1999. In contrast, the public sector’s employment share rose by 3½ percentage points to slightly above 18 percent in 1991, in part as the government attempted to improve labor market conditions directly by employing increasing numbers in the general government, as well as through expansion of public enterprise hiring. Thereafter, attempts to improve the public finances (with an eye to meeting the Maastricht criteria) led the government to impose in 1996 in a l-for-5 general government hiring rule (i.e., one hire for five departures); this rule did not apply to education, healthcare, and defense. In addition, protracted attempts to aid “ailing enterprises,” some of which had been nationalized in the wake of the 1970s oil price shocks, were wound up, resulting in the closure of numerous firms, while other overstaffed state-owned enterprises and banks, especially those facing prospective privatization, limited new hiring. These efforts have severely limited traditional employment sources for many, especially those with tertiary education.31

Figure 5.
Figure 5.

Greece: Employment Shares

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: National Statistical Service of Greece; and Fund staff calculations.

75. The share of self employment has steadily declined in recent decades, but still accounts for more than 40 percent of total employment, a figure much higher than in most other euro-area economies (although the share is also high in Italy).32 As a result, the average enterprise size in Greece, at 1.9 individuals, is far smaller than in the EU (7.2, see OECD, 1998a). The use of part-time employment in Greece remains limited, despite the progressive lifting of restrictions on its use in the 1990s, accounting for only 6 percent of employment in 1999, half the EU average rate. The share of females in part-time employment, about 10 percent of total female employment in 1999, was almost three times the male share. In addition, most part-time work may be “involuntary,” as a recent survey indicated that 70 percent of Greek part-time workers in Greece would prefer full-time employment (European Commission, 2000c). Temporary workers on fixed-term contracts constituted 7½ percent of total employment and 13 percent of dependent employment, similar to the EU average, with many seasonal workers (including those in the tourist sector) hired on such contracts.

76. Greece has been among the unfortunate group of economies experiencing a rising structural unemployment rate in the 1990s, with only two other OECD countries estimated to have higher rates (OECD Analytical Database). In addition, the composition of unemployment displays a number of features indicative of a poorly functioning labor market. Unemployment duration steadily worsened until the mid-1990s, with the majority of unemployed having lacked jobs for more than one year, worse than the EU average (Figure 6, and compared with only 8 percent of total unemployed in the United States). The majority of the unemployed have no previous employment experience, a rate 2½ times that in the EU and three times the euro-area rate, reflecting the extreme difficulties facing particularly the young and women. In fact, the male youth (15–24 years) unemployment rate is almost 25 percent, while for females the youth rate exceeds 40 percent (Figure 4).

Figure 6.
Figure 6.

Greece: Unemployment Characteristics

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: International Labor Office (1999), and Labor Statistics Database;OECD (1999b); and Fund staff calculations.

77. Another indication of Greece’s poor labor market performance is the absence of the well-known pattern of unemployment rates declining among cohorts with higher levels of educational attainment. Rather, Greek unemployment rates display a “hump” pattern, with higher rates for women aged 25–64 with upper secondary educations compared with those with less than upper secondary educations (Table 1). Among men of the same age group, the unemployment rate increased steadily with rising educational levels, with the exception of those with academic tertiary-level educations or beyond. This pattern is even more pronounced among youth aged 15–29, with unemployment rates among those with tertiary educations double the rates of those with less than secondary educations, and triple those among all OECD economies on average (Table 2). Among the explanations for this pattern is the continued high share of agricultural employment in Greece, providing employment opportunities to those with lower educational attainment. But one would expect that the pattern would be less pronounced for youth, as fewer individuals are entering this declining sector. Thus, reasons behind this humped unemployment rate pattern may rather be due to a poorly performing educational system and to labor market segmentation between “insiders” and “outsiders” (discussed further below).

Table 1.

Unemployment Rates by Level of Educational Attainment and Gender for Populations 25 to 64 and 30 to 44 Years of Age (1998)

article image
Source: OECD (2000a).

Year of reference, 1997.

Tertiary type B includes some post-secondary graduates.

Included in tertiary-type A.

Table 2.

Youth Unemployment Rates by Level of Educational Attainment and Age Group (1998)

article image
Source: OECD (2000a)

Year of reference, 1997.

Included in tertiary-type A.

78. The regional dimension of unemployment is of less importance in Greece than in some other EU economies. Unemployment rates generally tend to be lower in the islands (ranging from 8.1 percent of the labor force on Crete to 11.6 percent in the Southern Aegean in 1999, with a simple average of 9.8 percent) than on the mainland (from 7.8 percent in the Peloponnese to 14.6 percent in Western Macedonia, with an average of 12.3 percent). Greek regional unemployment rate disparities are generally lower than in other EU economies, as the coefficient of variation, at 24.7 in 1997, was less than those in 7 other countries, while higher than in 5 other countries (with the coefficients of variations ranging from 4.1 in Ireland to 61 in Italy, OECD 2000b).33 The coefficient of variation fell to 18.3 in 1999, although as the regrettable result of relatively larger increases in unemployment rates in those regions that had previously had relatively low rates. Higher unemployment rates also tend to be more of an urban phenomenon, with a rate of 13.7 percent in urban areas in 1999, compared with 10.1 percent in semi-urban areas, and only 7.4 percent in rural areas.

C. Labor Market Institutions

79. Greece’s labor market institutions have undergone extensive changes in the past two decades. While the wage formation system has been broadly liberalized, it remains rather centralized, with the government maintaining a large, although generally indirect, role in the private-sector wage formation process. Moreover, a legacy of previous governments’ more direct influence on wage formation, namely a relatively compressed wage structure, continues to influence labor market outcomes, especially when combined with a sizable increase in labor taxation, to the detriment of new labor market entrants’ and the less-skilled’s employment prospects. Greece also has among the most strict employment protection legislation among OECD economies. While this could have beneficial effects regarding in-work training, and there is little evidence internationally that this directly raises the unemployment rate, there is evidence that it is associated with reduced labor market flows and longer unemployment spells, and that the costs of employment protection afforded to prime-age males is borne by prime-age females, the youth, and older workers. Greece has a very limited unemployment benefit system, which minimizes the negative labor market incentives that have been found to have been important in increasing unemployment levels and duration in other EU economies. However, Greece also has had relatively limited active labor market programs, lending little support to those attempting the transition from education to employment, or in retraining workers. Attention must also be paid to other social and economic aspects affecting the labor market, including product market competition, entrepreneurship, innovation and technology, and the educational system. Unfortunately, Greece is found to have extensive product market regulations, stifling red tape, a barely nascent “new economy,” and weaknesses in its educational system.

Wage formation

80. Prior to the early 1990 reforms, the government was actively involved in the wage formation process. Following the fall of the military junta in 1974, and in reaction to wage suppression that had occurred under that regime, subsequent governments effectively imposed sizable increases in the minimum wage that exceeded productivity growth. In addition, wage increases in both the public and private sectors in the 1980s were determined under an automatic indexation system linked to inflation, but with less than proportional indexation for higher incomes. However, this system was overridden at times (in response both to economic considerations, such as the adjustment program in 1986–87, as well as to the electoral cycle). In addition to creating wide swings in real wages, this intervention resulted in a pattern in which the ratio of minimum wages to average earnings was relatively high, and relative wages were compressed (Figure 7).34

Figure 7.
Figure 7.

Greece: Real Wages and Productivity

(Natural logs, In (1981) = 0)

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: Bank of Greece (1998); National Statistical Service of Greece; OECD Analytical Database; and Fund staff calculations.

81. The regressive automatic wage indexation system was abolished in 1991 (although a system for compensating private sector workers for higher-than-expected inflation now exists). The collective bargaining system was also decentralized and broadened. Greece now has a multi-tiered system of collective agreements, with the broader agreements, beginning with the national collective agreement between labor and employers concerning the private sector minimum wage, setting effective floors for subsequent sectoral or enterprise agreements. As a result, some measures indicate that wage compression has been somewhat mitigated in the 1990s (Figure 8).

Figure 8.
Figure 8.

Greece: Wage Dispersion

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: Bank of Greece; National Statistical Service of Greece; and Fund staff calculations.

Tax wedge

82. The role of taxes on the level of unemployment is somewhat controversial (Nickell, 1997).35 A comparison of the tax wedge in Greece with its EU and euro-area partners indicates that the total wedge is almost identical to its partners’ average wedge, and had changed little in the mid-1990s (Table 3).36 It is, however, higher than rates in other lesser-advanced partners (Ireland, Portugal, and Spain), and significantly higher than those, for example, in the United States. From a broader historical perspective, the total tax wedge (the sum of effective average labor and consumption tax rates) has increased markedly, with higher labor as well as consumption taxes (Figure 9). This reflects increased payroll taxes in the context of the 1990–92 social security reforms, personal income tax bracket creep from a failure to adjust tax brackets fully for inflation, and the introduction of a value-added tax in 1987 (with an increase in the basic rate to 18 percent in 1998).37 While the effects of bracket creep most likely have had little impact on those earning minimum wages (which fall into the zero-tax bracket), the boost in payroll taxes could have had deleterious effects on labor demand, especially for the unskilled and new labor market entrants.38

Table 3.

Tax Wedge and Effective Tax Rates on Employed Labor, 1998 1/

article image
Source: European Commission (1999).

All tax rates have been calculated on the labor of wage and salary earners, thus excluding the self-employed.

Tax wedge = (I - (I - (SSC + income taxes)/100)*(1 - consumption effective tax rate/100))*100.

SSC = social security contributions: payroll taxes are included in employees’ SSC; both rates are related to gross labor costs.

Income tax = personal income tax*(gross labor costs - SSC)*100/gross labor costs.

Consumption effective tax rate = indirect taxes*100/(private consumption + public consumption - gross labor costs of public sector employees).

Figure 9.
Figure 9.

Greece: Effective Average Tax Rates

(In percent)

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Source: Lutz (1997).

Employment protection

83. Employment protection legislation (EPL) in Greece is among the strictest from an international perspective, with only Portugal and Turkey considered to have stricter systems (Tables 47). It is interesting to note that on the basis of these summary indicators, Greece’s system has not materially changed since the late 1980s. In contrast, ten other economies liberalized their systems, especially Sweden, Denmark, Italy, and Germany, and mostly regarding temporary employment. 39 It is also notable that among the regional groupings, employment protection remains clearly the strictest in Southern European economies, despite the liberalization efforts in Portugal, Italy, and Spain.40 In examining the composition of the summary indicator, Greece’s EPL strictness for regular employment is about average, the sixteenth strictest of 27 economies. It has the same relative ranking regarding its regulation of collective dismissals. With regard to the regulations regarding temporary employment, however, Greece’s were among the strictest. Temporary work agencies were, until late 1998, illegal.

Table 4.

Summary Indicators of the Strictness of Employment Protection Legislation 1/

article image
Source: OECD (1999c)

1/ Higher values represent stricter regulations.

2/ Average of indicators for regular contracts and temporary contracts.

3/ Weighted average of indicators for regular contracts, temporary contracts, and collective dismissals.

Table 5.

Indicators of the Strictness of Employment Protection for Regular Employment 1/

article image
Source: OECD (1999c).

The summary scores can range from 0 to 6 with higher values representing stricter regulation.

Table 6.

Regulation of Temporary Employment 1/

article image
Source: OECD (1999c).

Higher values represent stricter regulation.

Table 7.

Regulation of Collective Dismissal, Late 1990s

(Requirements over and above those applying to individual dismissal)

article image
Source: OECD (1999c).

The score is 0 if there are no special regulations on collective dismissal; 1 if regulations apply from 50 dismissals upward;. 2 if they apply from 20 onward; 3 if they start at 10 dismissals; and 3 if regulations start to apply at below 10 dismissals.

There can be notification requirements to employee representatives/works councils, and to government authorities such as public employment offices. Countries are scored according to whether there are additional notification requirements on top of those requirements applying to individual redundancy dismissal. The score is 0 if there are no additional requirements: 1 if one more actor; and 2 if two more actors need to be notified.

This column lists delays required on top of delays before the start of notice for economic redundancy listed under Table 5. Averages are taken if separate delays apply to different types of situations.

This column refers to whether there are additional severance pay requirements in case of collective dismissal and whether social compensation plans (detailing measures for redeployment, retraining, outplacement or severance pay) are obligatory or common practice. The score is 2 if both requirements apply.

The summary scores can range from 0 to 6, with higher values representing stricter regulation.

84. The economic impact of strict EPL, especially for the unemployment rate, is mixed.41 As intended, it reduces the separation rate from employment into unemployment, but will also reduce the flow out of unemployment into work (which should apply to new school leavers as well), as firms are cautious about new hiring (Nickell 1997). This appears to be broadly consistent with the Greek case, which has the lowest rates of inflow and outflow from those unemployed among OECD members, as well as one of the lowest shares of dismissed workers in total unemployment (OECD 1996). Information regarding the degree to which EPL is a binding constraint in Greece is unclear, but it appears to be related to the business cycle, although EPL would not apply to Greece’s large share of self-employed (and clearly, not to the underground economy either).42 It has been suggested that EPL, by reinforcing job security, may enhance productivity performance, as workers are more willing to help in improving production processes, and as long-term employment relations may make employers more willing to provide training (Scarpetta, 1998). Unfortunately, these potential benefits have largely not materialized in Greece. Until recently, the level of factor productivity was largely stagnant (Lutz, 1998; and Figure 7). In addition, Greece has the lowest participation rate in career- or job-related training among OECD economies (OECD, 1999c). Strict EPL has been found also to be related to lower short-term unemployment and greater long-term unemployment, a situation borne out in Greece (Figure 6). But the overall effect on the level of unemployment is unclear, as the fewer flows into unemployment are potentially offset by longer durations.

Unemployment benefits

85. In contrast to the strict legislation protecting the employment positions of traditionally prime-age males in Greece, the level of assistance provided to the unemployed is quite limited. The average net replacement rate in Greece is roughly one-half of the average production worker’s net income, compared with 65–75 percent on average for the EU and euro area (Table 8). In addition, the duration of unemployment benefits, totaling 12 months for most workers, is among the less generous.43 Accumulated empirical evidence suggests a clear relationship between the generosity of unemployment benefits (both in level and duration) and unemployment (Scarpetta, 1996; Nickell, 1997; Nickell and Layard, 1997; Elmeskov, Martin, and Scarpetta, 1998; Blanchard and Wolfers, 2000). Thus, it would appear that the deleterious effect for job-search incentives from overly generous benefits is largely absent in Greece.

Table 8.

Net Replacement Rates of the Unemployed, 1997 1/

article image
Source: European Commission (1999).

Net replacement rate for each category are the averages of the net replacement rates for three family types: single person, couple without children, and couple with two children.

APW = average production worker.

The 60th month’s net replacement rate includes possible topping up of social assistance in addition to unemployment benefits whereas the 1st month’s rate includes the unemployment benefit alone.

Weighted by nominal GDP; EU-15 in 1995 excluding Greece.

Active labor market policies

86. As with unemployment benefits, the level of assistance provided to those seeking to obtain or return to employment is limited in Greece. Greece is among the lowest spenders in the EU and euro area on active labor market programs (ALMPs), with considerably lower expenditures in all areas except for youth measures (Table 9.)44 Similarly, the macroeconometric evidence regarding ALMPs generally suggests that such policies are effective in reducing the overall unemployment rate, and especially the long-term unemployment rate (Nickell, 1997; Nickell and Layard, 1997; Blanchard and Wolfers, 2000).45 It would thus appear that there are considerable opportunities to reduce unemployment through targeted measures.

Table 9.

Benefit Duration and Job Availability Requirements

article image
Source: European Commission (1999).

For a 40-year old single worker with a long employment history.

The index is a weighted average of points from 1 to 5 dedicated to eight categories of job availability criteria. Five points are given to the maximum strictness.

Depends on the length of the employment history.

Depends on the age of the worker becoming unemployed.

Other influences

87. In addition to labor market institutions, it has become widely recognized that other social and economic factors have significant influences on labor market behavior. Included among those are the degree of product market competition, the scope for entrepreneurship, the government’s and private sector’s involvement in innovation and technological activities, and the quality of a nation’s education and training infrastructure. As the first two topics are covered in the staff report (Box 3), and the third topic in Chapter IV, little additional will be said here.

88. Greece’s educational system has been challenged to adapt to a rapidly changing economic and social environment, lest the skills obtained by recent graduates become misaligned with those demanded by employers, resulting in protracted spells of unemployment among new labor market entrants. Greece’s eighth grade math achievement scores are significantly below OECD, EU, and euro-area averages. From an international perspective, the educational attainment of Greece’s working age population is quite bifurcated, as those aged 25–64 with only lower secondary educations accounted for 54 percent of the labor force in 1998, compared with a 38 percent simple average for OECD economies.46 In contrast, the share of the population with tertiary-type A (i.e., academic-oriented collegiate educations) and advanced degrees was 11.3 percent, close to the 14.1 percent average. The gap in the percentage of the population aged 25–34 that has attained at least upper secondary educations compared with the OECD average has narrowed significantly, although the percentage of 17-year-olds participating in secondary education still lags the average.

89. Several aspects of secondary and tertiary-level education in Greece suggest the potential for labor market mismatches. Fifty-seven percent of those aged 20 are enrolled in tertiary education, compared with an OECD average 25 percent enrollment rate. This reflects the high social status afforded to university graduates.47 As a result, “[secondary education] no longer functions as an independent and self-contained school, but has been transformed into a preparatory level, like a waiting vestibule, for the universities” (OECD 1997, p. 37). In contrast, the share of students involved in pre- and vocational programs is half the OECD average. Recalling the “hump-shaped” unemployment pattern in Greece, this is suggestive of a possible relative shortage of vocationally trained employees. Greece’s public expenditures on education, at 3.5 percent of GDP, lags significantly behind the 4.8 percent OECD average. In contrast, the level of private expenditures, at 1.4 percent of GDP, exceeds the average, despite the minimal role of pre-tertiary private educational institutions (as well as the constitutional monopoly on public university-level educational institutions), reflecting in part expenditures on “cramming schools” (frontisteria) for those preparing for university entrance examinations. The public monopoly on university education, as well as severe overcrowding at the tertiary level (with student/teacher ratios almost twice the OECD average, in part reflecting the second longest average duration of tertiary education at six years compared with a four-year average), has resulted in Greece being among the countries with the largest proportion of students studying abroad. It also appears that the formation of educational-employment linkages is seriously lacking in Greece. It has—along with Italy, Portugal, and Turkey—the lowest share of young students who are simultaneously employed (at about ¾ percent for 15–19 year olds compared with a 14½ percent OECD country average). This lack of familiarity with working life requirements appears to be correlated with subsequent spells of youth unemployment (Figure 10).48 Thus, it would appear that the present educational system, which progressively became geared toward producing generalist university graduates destined for public sector employment, is not well-suited for meeting today’s private sector labor demand.

Figure 10.
Figure 10.

Percentage of 15-19 year-olds in Education who are Employed and Unemployment to Population Ratio of 20-24 year-olds not in Education, 1998 1/

Percentage of 15-19 year-olds in Education who are Employed and Unemployment to Population Ratio of 20-24 year~olds not in Education, 1998 1/

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: OECD (2000a); and Fund staff calculations.1/ Data for Denmark, Greece and the United States are for 1997.

D. The Interaction of Economic Shocks and Institutions

90. Greece’s high unemployment rate has resulted from employment growth failing to absorb a rapidly growing labor force. What are the forces that have hindered the creation of more employment? This section analyzes structural demand and supply side factors that have influenced labor market developments.

91. How far do real wage developments go toward explaining the observed patterns of unemployment? Until the mid-1990s, business sector real wages were broadly constant in Greece, while they had generally increased in the other economies considered (Portugal, Spain, Italy and the euro area, Figure 11).49 Subsequently, Greek real wages grew rapidly, such that the increase during the 1990s exceeded that in the euro-area average (recall, though, that the levels of real wages in Greece, expressed in a common currency, and productivity are still well below euro-area averages). Greece’s real wages increased in line with those in the euro area in the 1980s, and both had broadly similar unemployment rate developments. Greece experienced wide swings in the real wage in the 1990s; in contrast, while the unemployment rate rose steadily, euro-area real wages, with the exception of a dip in 1991 (reflecting German reunification), continued their upward drift, while the unemployment rate initially rose further before it eventually began to decline. Thus, it would seem that factors beyond real wage developments (which here have implicitly been assumed to be shifts in labor supply) have accounted for contrasting developments in Greece and the euro area in the 1990s and some other countries in Southern Europe.

Figure 11.
Figure 11.

Real Wages and Unemployment

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: OECD Analytical Database; and Fund staff calculations.

92. Clearly real wage growth can be the result of other factors, such as productivity growth, or changes in labor demand, which could result in different real wage/unemployment rate patterns. In order to examine more fully the relative roles of underlying, or structural, labor demand and supply shifts, this paper uses the framework in Blanchard (1998), in which he explores the effects that shifts in these schedules on relative factor prices, factor income shares, and labor/capital ratios. For simplicity, the analysis is conducted in a static framework, examining shifts of the labor supply or demand curves, although it can easily be applied to a growing economy if one thinks of shifts in the supply and demand schedules relative to their movements along a balanced growth path.

93. The labor demand schedule provides for the standard inverse relationship between real wages and the amount of labor demanded, except the wage is expressed in Harrod-neutral (i.e., labor-augmenting) efficiency-adjusted terms.50 The demand curve can shift from changes, for example, in union power, which may alter the amount of featherbedding that they are able to impose, or from movements in technology that are biased either toward or against capital. The shifts in the labor demand curve over time are given by:

ΔLtd=Δ[log(wt/at)+1σlog(atntyt)](1)

94. Conveniently, for σ= 1, when a CES-type production function simplifies to a Cobb-Douglas functional form (where σ is the constant elasticity of substitution), the evolution of labor demand is given by shifts in the natural logarithm of the labor share of income.

95. Labor supply is proposed to be a positive relation between the wage (in efficiency units) and the employment (which, for a fixed labor force, is equivalent to a negative relationship with the unemployment rate).51 Real wages are able to grow at the rate of labor-augmenting technological change with no alteration in unemployment (while for a growing economy, the schedule would be shifting out over time along a balanced growth path to maintain the unemployment rate/real effective wage rate ratio). The relationship can be expressed as:

w/a=-βU+Z,(2)

where U stands for the unemployment rate. Changes in Z can be thought of as shifts of the supply function. Among the possible factors responsible for this are changes in wage bargaining structures, unemployment benefits, and EPL. An increase in Z suggests that workers demand a higher efficiency-adjusted wage for a given level of unemployment (alternatively, an increase in the level of unemployment, if considered to be of “outsiders” that have no influence on wage-setting behavior, would result in a leftward shift in the labor supply schedule).52 The following expression characterizes the evolution of labor supply Ls over time:

ΔLts=-ΔZt=-Δ[wtat+βUt](1)

96. Construction of both the labor demand and supply curves requires a time series for the real efficiency wage (Figure 12).53 In all three countries considered for which data are available, real wages increased by 15–20 percent during the last two decades, although they dipped in Greece in the early 1990s in the context of the government’s medium-term adjustment program to tackle growing domestic and external imbalances, and the beginning of enterprise restructuring. In marked contrast to developments in Italy and Spain, where the effective real wage declined over this period, it increased in Greece. This was the result of Greece’s relatively poor pace of technological improvement, averaging only ⅔ of 1 percent per year, compared with 1½ percent in Italy and Spain.

Figure 12.
Figure 12.

Real Wages, Productivity, and Effective Real Wages

(Natura1ogs, In(l981) = 0)

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: OECD Analytical Database; and Fund staff calculations.

97. The calculations suggest that labor demand shifted downward in all the countries considered (Figure 13).54 This is consistent with either a reduction in union’s relative power (for example, a drop in featherbedding), or a technological bias toward capital.55 These motivations have observationally equivalent results in this framework. For a given capital stock, the declining demand for labor initially reduces the effective wage and increases unemployment. Adjustment costs (including EPL) may make this a protracted process. The reduction in employment raises the profit rate and the capital share, with the former in turn inducing capital accumulation. The increasing capital/labor ratio boosts labor’s marginal product and increases employment. As a first-order approximation, the process ends with employment and effective wages returning to their initial levels, while the labor/capital ratio is permanently lower. Given an unchanged profit/wage ratio, but a lower labor/capital ratio, capital’s income share also permanently widens.

Figure 13.
Figure 13.

Labor Demand Developments

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: OECD Analytical Database; and Fund staff calculations.

98. In examining developments regarding capital’s income shares and labor/capital ratios, it would appear that the demand shift story goes a long way toward explaining developments in Italy (Decressin, 2000) but only partly in Greece and Spain (Figure 14). Italy’s capital income share has risen fairly steadily and its labor/capital ratio fell steadily. While Greece’s and Spain’s’ labor/capital ratios fell as well, Greece’s capital income share has changed the least over the last two decades. Moreover, it has been broadly constant, or even declining, since 1994. Spain’s capital income share rose sharply in the 1980s, when its labor demand was falling sharply, but it also has been somewhat more stable in the 1990s than has Italy’s, and its labor/capital ratio fell more dramatically than in the other economies.

Figure 14.
Figure 14.

Capital’s Income Shares and Labor/Capital Ratios

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: OECD Analytical Database; and Fund staff calculations.

99. There have also been sizable labor supply shifts in Spain and, to a lesser extent, in Greece, as measured by the Blanchard framework (Figure 15). In contrast to a reduction in demand for labor having a temporary, albeit potentially protracted, impact on effective real wages and unemployment, a shift in the supply of labor permanently raises unemployment. As the capital stock and technology are fixed initially, the declining labor supply results first in an increase in the real effective wage and an offsetting decline in profits and capital’s income share. The change in relative factor prices induces a reduction in employment and a partial reduction in the wage and rebound in profits and capital’s income share. The decline in capital induces a backward shift in the labor demand schedule, with the results discussed above. In the end (and assuming for simplicity a unitary elasticity of substitution), unemployment is higher and the capital stock is lower, while the labor/capital ratio and the wage and profit rate are unaffected.

Figure 15.
Figure 15.

Labor Supply Developments

Citation: IMF Staff Country Reports 2001, 057; 10.5089/9781451816136.002.A003

Sources: OECD Analytical Database; and Fund staff calculations.

100. Without pushing the analysis too far, it is possible to see the effects of shifts in both labor demand and supply schedules in Greece’s recent economic history. First, the sharp increases in minimum wages in the early 1980s, which were partly echoed in general wage increases, acted to shift the labor supply schedule leftward. This reduced capital’s income share and boosted unemployment. The latter was kept from further increasing, as discussed earlier, in part through direct government hiring and additional public enterprise employment. This, combined with other direct intervention in economic activity (including price and credit controls), and a strong political business cycle, resulted in stop-go macroeconomic policies, with rising labor demand in the early 1980s followed by reduced demand in the latter half of the decade. The short-lived mid-1980s adjustment program is reflected in a small drop in real wages in the 1986–87, before rebounding again for the remainder of the decade. Demekas and Kontolemis (1996) have argued that the government worsened labor market behavior in the 1980s, as the expansion in the number of relatively desirable life-time government jobs and an increase in public/private relative wages depressed private sector employment, boosted workers’ reservation wages, thereby contributing to the rise in unemployment.

101. The 1990s have seen a marked leftward shift in the labor demand schedule, as industry initiated restructuring efforts and the government began reducing public enterprise employment. At the same time, the supply schedule initially shifted rightward, as labor accepted lower wages for reasons of macroeconomic stability as well as in response to reforms in the labor market bargaining system. While these two shifting schedules should unambiguously reduce the effective wage, the impact on unemployment is ambiguous. In the event, the labor market was confronted with a resumed increase in female participation rates (Figure 4) and the economy absorbed large immigrant inflows. In retrospect, the fall in real wages, while necessary, was insufficient to price in the additional potential employees into jobs. Subsequently, real wage losses have been more than recouped, with gains again outpacing resumed technological progress. This is reflected in a leftward shift in the labor supply schedule by a cumulative 12 percent over 1991–99, with real effective wages rising in the face of increasing unemployment.

102. Why have real wages failed to adjust sufficiently for the labor market to employ the growing labor force? A number of possible, and possibly interrelated, reasons exist. First, and despite sizable short-run fluctuations, real wages have been found to be rather rigid in Greece (Halikias and Sobczak, 1998). This pattern is reinforced by the existence of private sector inflation catch-up clauses. It also is reinforced by the pattern of wage setting in Greece. The government maintains an influence over general wage developments, given its role as employer of 20 percent of all dependent employees. Moreover, wage increases for civil servants are announced with the annual budget, prior to the private sector agreements, and are considered to set a floor for the national collective minimum wage agreement. Greece’s wage bargaining system, with the national minimum wage agreement followed by narrower agreements at the industry and possibly enterprise levels, generally results in wage drift, with better performing industries and enterprises offering higher awards. Calmfors and Driffill (1988) have stressed that the relationship between the centralization of wage bargaining and wage outcomes is “humped,” with either highly centralized or completely decentralized bargaining systems generating the best results regarding structural unemployment.56 This is because the former internalizes economy-wide labor market pressures, while the latter results in atomistic agreements, each of which is too small to affect other agreements, or to have any economy-wide wage or employment effects. Greece, as in many other EU economies, has a system somewhere in the middle, where each industry-wide agreement is large enough to influence overall employment and wage levels, while at the same time there are enough separate agreements that it is difficult to coordinate them. Finally, it is curious why the national minimum wage agreement, one that is both centralized and coordinated, has not been flexible enough to respond to the growing unemployment rate. It would appear that Greece faces severe labor market segmentation, where wages are set to protect employment prospects of prime-age male workers (the insiders), with little attention is given to those who have yet to establish job histories (youth), have low skills, or hold nontraditional employment arrangements (largely female part-time workers).

E. Reform Efforts and Potential Extensions

103. Greece’s labor market problems, especially when contrasted with the euro area’s improving unemployment picture, are glaring. The government has adopted reforms to address a number of weaknesses. While for the most part they are steps in the right direction, they may be insufficient by themselves to significantly improve the labor market’s functioning, as they fail to address some of the key structural weaknesses discussed above.

104. The government adopted a law in 1998 to legalize the large number of illegal immigrants. By mid-2000, 380,000 residency permits (“white cards”) had been requested and 223,000 employment permits (“green cards”) had been asked for (OECD, forthcoming). One requirement for obtaining a green card was joining the social security system, which, given the relative youthfulness of economic immigrants, should improve the pension system’s financial footing.57

105. As part of an agreement between social partners in November 1997 on a medium-term strategy for reforming the labor market, the government adopted a law in August 1998, containing five main innovations: a calculation of working time over longer periods to reduce overtime and facilitate seasonal employment needs; “opt-out” options from sectoral and enterprise-level wage pacts (while imposing the national collective minimum wage increase) in pre-determined geographic areas with high unemployment or industrial decline; the promotion of part-time employment by elimination of barriers to its use; the authorization of private job placement agencies; and the regulation of informal forms of employment (e.g., piecework). The results of these efforts to date have been limited.58 This was in part due to lags in implementing decrees and directives until 1999. In addition, the measure allowing for the annualization of working hours was little used, as the law as adopted required enterprise level collective agreements (which do not exist for most small enterprises, and which predominate in Greece), rather than by the initial proposal to have a unilateral employers’ decision.

106. In the context of annual budgets, the government has adopted a number of tax measures intended to reduce hiring costs for employers and increase take-home pay for some workers. In September 1999 it established tax incentives for new personnel recruited between mid-November 1999 and end-December 2000, with 50 percent of employers’ social insurance contributions being deducted for a two-year period from net income (and thus reducing the corporate income tax base) for those hiring firms that did not have redundancies over the same period. The budget also increased unemployment benefits by 10 percent, and provided free medical care to all registered unemployed. In the spring of 2000, the government took over financial responsibility for employee-paid pension contributions for minimum wage recipients, equivalent to 6⅔ percent of earnings, and boosting their disposable incomes by 8 percent. It was estimated that some 470,000 individuals would be affected by this measure, including some part-time workers. The 2001 budget contained an indexation of the personal income tax brackets of 5 percent, and a reduction in the top marginal tax rate of 2½ percentage points (to 42½ percent), while a similar reduction in the top rate was announced for 2002. The government has also begun efforts to fundamentally reform the Greek tax system, broadening tax bases and reducing rates, to be implemented in three years.

107. The government has bolstered its ALMPs in the last few years, and has begun to address weaknesses in the educational system. The Ministry of Labor is reorganizing the public employment service (PES), with a business plan under preparation. The PES had opened 49 employment promotion centers by late 2000, with a doubling of this number anticipated by 2006. These centers use an individual counseling approach in helping to provide employment. While private employment offices have been legalized since the 1998 reforms, the first one opened only in November 2000, and its scope was limited to a list of prescribed occupations, the result of strong union opposition. Turning to ALMPs, some 54,000 subsidized employment positions had been created in the first nine months of 2000 (against an annual target of 75,000, equivalent to some 1.7 percent of the labor force), while other employment programs (including those for freelances and unemployed graduates) found places for a further 23,000 individuals. It also intended to create 75,000 new training positions. In addition, a system for monitoring and evaluating programs was under preparation, an activity urgently needed to assess the strengths and weaknesses of the various programs, and to facilitate continued improvements. The Ministry of Education and Religion adopted in 1998 educational reforms intended, inter alia, to raise the status of technical and vocational education, improve guidance and counseling services, encourage closer links between the educational system and the labor market, and expand the number of tertiary education positions (while retaining the public monopoly).

108. The new government, elected in the spring of 2000, stated that increasing employment was one of its top priorities. New reform proposals to improve labor market flexibility were discussed with social partners, but faced strong resistance from both employers and unions. Nevertheless, parliament adopted in late December 2000 a package of measures intended to boost employment, reduce the use of overtime, ease firing restrictions, and reduce long-term unemployment. Specifically the package includes a reduction in the maximum workweek from 48 hours to 43 hours, with the amount of overtime worked at the employers’ discretion reduced from 8 hours to 3 hours, and with a shift in its remuneration premium from 25 percent to 50 percent, and an increase in the fine for “illegal” overtime from 100 percent to 150 percent; alternately, an incentive was provided to annualize working hours (allowing for more working hours during peak periods but without overtime premia, offset by fewer hours worked during slack periods) by reducing the average workweek to 38 hours, which would require employee approval; reducing labor costs by cutting the employer-paid social security contributions by 2 percentage points for those with monthly earnings of less than Dr 200,000 (available from April 1,2001 through 2003); efforts to boost part-time employment by increasing the minimum wage for those working fewer than four hours a day by 7.5 percent; and having the public employment service pay for up to one year monthly benefits equivalent to Dr 30,000 to previously long-term unemployed who are hired part-time for at least four hours daily; an easing of firing restrictions for medium-size firms (while tightening them for small firms);59 the state paying the social insurance contributions of the long-term unemployed close to retirement age; reducing the retirement age for those having worked in arduous and unhealthy occupations; and the state providing a rent subsidy for up to two years for the long-term unemployed.

109. While these laws go some way in improving labor market flexibility, they have a number of shortcomings. For example, the first measure allowing for a profit tax credit for new hires was only for a limited time period and was not of use to firms without profits (presumably including new firms starting up). Second, the possibilities to introduce more working time flexibility was obtained, but only at the expense of rising labor costs, either directly through higher overtime premia, or indirectly through a reduction in the number of hours worked (from on average 40 a week to 38) without overtime payments. Third, the reforms in firing rules eliminates the much-discussed “kink” that existed previously, but is now in fact tighter than before for small firms, and has not changed for the largest enterprises.

110. In addition, the recent initiatives fail to address some of the key structural weaknesses described above, and therefore are unlikely to achieve a major reduction in unemployment. Notably absent are measures addressing the root causes of exceptionally high unemployment rate among new labor market entrants—an area where a comprehensive strategy, involving steps by the social partners as well as the state, could be considered (and as was recommended in the staff report). Increased wage differentiation for first-time job seekers(half of all unemployed) could be a useful additional measure to help reduce employer costs for those with initially low productivity levels. While the debate on the effects of minimum wages on employment is heated, a review of recent work and independent estimates by the OECD suggests that higher minimum wages adversely affect teenage employment(OECD 1998b).60 There are valid concerns that the income derived from subminimum wages would not provide an adequate living standard. Even if this were the case, it would be possible to design a tax-based supplemental income scheme, similar to the earned income tax credit system in place in the United States.61

111. Another reason that the labor market has been incapable of employing the rapidly growing labor force is due to the real wage rigidity resulting from the wage bargaining system. With euro-area participation, expectations regarding generally low and stable inflation should spread, and accordingly the need for the inflation catch-up clause could disappear. The multi-tiered wage-setting process also appears to impart an upward bias to wages, and consideration could be given to scrapping the industry-wide tier of bargaining. Measures could also be found to allow smaller firms to reach “opt-out” agreements in high unemployment/declining industry areas.

112. The recent reforms to the firing restrictions were at best marginal, and for smaller firms, retrograde. Given the economy’s present cyclical position, the restrictions do not appear to be binding, although they still influence the hiring decisions of some employers. Nevertheless, they still have adverse influences on more marginal labor market participants. Thus, further liberalization could be considered.

113. The authorities are making inroads in improving their ALMPs. However, the amount of spending on these programs remains well below international averages, and the bulk of job placement and training activities remains under the purview of the state. Emphasis could be shifted from subsidized employment positions to training programs. The latter would benefit from comprehensive input by social partners to ensure their relevance to labor market needs.

114. The potential benefits from the proposed educational reforms should not be underestimated. In light of the authorities’ candid self-assessment (OECD 1997) and reviewers’ comments and recommendations, the 1998 educational reform program, when implemented, should improve the efficiency and economic relevance of the educational system.62 In particular, restructuring the secondary education system (especially for technical and vocational education) should help establish closer links with the labor market. In addition, large benefits would derive from measures to facilitate life-long learning and meet evolving labor market needs.

115. Greece’s labor market structure reflects both economic and political developments. While all agree that reforms are necessary to improve its functioning, almost all reforms would result in at least transitional hardships for particular segments of society. Thus, recent theoretical and empirical analyses of successful labor market reforms stress the need for a broad package of comprehensive reforms, which may help to compensate losers from individual policy measures and help make reforms more politically acceptable (Coe and Snower, 1997; Orszag and Snower, 1998; and Elmeskov, Martin, and Scarpetta, 1998).

Table 10.

Greece: Public Expenditures on Active Labor Market Programs, International Comparisons

(in percent of GDP)

article image
Sources: OECD (1999c): and Fund staff calculations.

Unweighted averages

Derivation of the Demand Schedule

116. This appendix derives the structural labor demand schedule of Blanchard (1998).

117. Assume that production is characterized by a constant-returns-to-scale CES-type production function with Harrod-neutral technological progress:

y=A[α(an)σ-1σ+(1-α)kσ-1σ]σσ-1,(1)