FYR Macedonia

Labor Market Profile 2004-2007

A Policy Note

June 27, 2008

Human Development Sector Unit

EUROPE AND CENTRAL ASIA

Document of the World Bank

CURRENCY EQUIVALENTS

(Exchange Rate Effective)

Currency Unit / =
US$ / =

Fiscal Year

July 1, 2007 / – / June 30, 2008

ABBREVIATIONS AND ACRONYMS

Vice President: / Shigeo Katsu
Country Director: / Jane Armitage
Sector Director: / Tamar Manuelyan Atinc
Task Team Leader: / Gordon Betcherman


FYR Macedonia Labor Market Profile 2004-2007

CONTENTS

Executive Summary 1

Labor Force Participation 7

Employment 12

Unemployment 23

Flows And Labor Market Dynamics 30

References 33

2

Executive Summary[1]

The employment situation in FYR Macedonia has improved over the last 4 years, but still the country’s labor market faces several challenges. This analytical note provides an employment profile for the working age population using the 2004-2007 Labor Force Surveys (LFS). This note is primarily descriptive. Results in this note indicate that labor market outcomes improved in the 2004-2007 period, with some increases in employment and labor force participation, and reductions in unemployment. However, labor market indicators are still far from the levels of the European Union (EU). Much of the recent progress has been achieved in relatively low-skilled, low-quality jobs. The new national employment strategy sets the goal to raise the employment rate to 48 percent in 2010. While reaching a level of 48 percent by 2010 remains an ambitious goal (employment rates reached approximately 40 percent in 2007), increasing the level of overall employment rates is only the starting point for a series of challenges in the country’s labor market.

Basic labor market indicators in FYR Macedonia are among the worst in Europe. FYR Macedonia is now an official candidate for joining the EU. The accession process constitutes an important challenge and opportunity for the country. The country needs to adapt its labor market to better capture the benefits of EU membership. Yet, the country has the highest level of unemployment in Europe with more than one third of its labor force in unemployment. Participation rates among the working age population are below those in the ECA region on average (65 percent in ECA vs. 61 percent in FYR Macedonia). This is due to lower than average labor force participation rates among females. The country’s employment rate (at 40 percent) is the lowest in Europe (averaging 60 to 65 percent) and far below the Lisbon target of 70 percent. While improvements of employment outcomes are cross-sectorial by nature as they relates to macroeconomic (investment climate and economic growth) and microeconomic (legislation and employment policy) performance, improving the conditions of the Macedonian labor market will not be an easy task.

Labor force participation rates among women are low for ECA standards. Female participation rates have remained flat since 2004 at 55-56 percent. FYR Macedonia is the second country in the ECA region (after Turkey) with the lowest female labor participation rate among the working age population. Among the reasons of female inactivity, household responsibilities are the most important. Approximately 55 percent of women who are not in the labor force are housewives. However, this proportion varies by educational level. In fact, 68 percent of women having primary or less education are out of the labor force due to housewife responsibilities, while the same share is only at 21 percent among women having higher or university education.

High levels of unemployment still remain the major single challenge in the FYR Macedonia labor market. Historically, FYR Macedonia was one of the regions with the highest unemployment rates in the former Yugoslavia. High rates of unemployment are a matter of policy concern, especially because of their negative social and economic repercussions. FYR Macedonia is the country with the highest unemployment rate in ECA at 37 percent. Unemployment rates in ECA vary from 5 to 20 percent. Montenegro and Bosnia and Herzegovina, display the highest unemployment rates (at 31 and 30 percent respectively) after FYR Macedonia.

More than half of all unemployed have never worked previously. According to 2006 LFS estimates, 55 percent of all unemployed individuals have never worked previously. The remaining 45 percent of the pool of unemployed have been laid off and/or quit their jobs. As expected, the fraction of first time job seekers decreases with age. In fact, while the share unemployed between 25 and 34 with no previous work experience is at 74 percent, the same fraction is at 30 percent (lower but still large) among individuals in prime working wage (35-54).

Ninety one percent of all unemployed who are not first-time seekers are long term unemployed. About 133 thousand individuals (mainly adults) could be classified as long-term unemployed, as defined by being in unemployment for a period longer than 1 year. Only 1 out of 10 unemployed (excluding first time seekers) has been unemployed less than 1 year, while roughly 5 out of every 10 unemployed individuals have been so for a period longer than 4 years. This is a very worrisome indicator. Generally, the longer individuals are unemployed, the less likely they are to find a job. Long term unemployment, albeit large across all age groups, seems to affect adults (35-64) more: while 3 out of 10 unemployed individuals between ages 15 and 34 have been unemployed for more than 4 years, the same indicator is at 6 out of 10 for individuals between ages 35 and 64.

However, estimates of unemployment rates are likely to be inflated because some jobless individuals who are not seeking for jobs actively are being classified as unemployed according to the standard definition used by the State Statistical Office (SSO). Results in this note indicate that a large fraction of individuals categorized as unemployed limit their job search to being registered in the employment office. As such, these individuals loose connection with the labor market and are more likely to be out of the labor force than unemployed.

By using a more strict definition of active job search, the unemployment rate would be much lower, albeit still high. A large fraction of all unemployed individuals rely on being registered in the Employment Office as a main method of job search. Active job search is one of the key requirements to be considered unemployed. The definition of active search includes individuals with different job search intensities. Results in this note indicate that about one third of all unemployed only search jobs through being registered in the Employment Office. This note provides estimates of “counterfactual” unemployment rates by age group using different definitions of active job search. By using more restrictive definitions, unemployment rates would decrease largely for all age groups:

·  Standard definition: Consists of individuals who are registered in the employment office, have tried to start an activity as self-employed, contacted private employment agencies, placed advertisement, answered advertisement, contacted employer directly, checked with relatives or friends, and participated in job testing/interviews.

·  Strict definition: Similar to the standard definition but excludes unemployed individuals who only seek jobs by being registered in the Employment Office.

·  Very strict definition: Includes individuals who placed an ad, answered an ad, contacted employer, and/or participated in job interviews.


Unemployment Under different definitions of “Active Search”

Source: Author’s estimates using 2004-2006 LFS data.

Flows across different employment statuses are scarce: employed individuals tend to remain employed, and unemployed individuals tend to remain unemployed for long periods of time. This phenomenon undermines the labor market potential for new entrants (mainly youth) who have to accept unpaid and/or very low quality jobs in order to acquire some work experience.

Trends in Labor Market Outcomes

There has been some improvement in labor market indicators over the period 2004-2007. Although labor market indicators are discouraging in the European context, important progress has been achieved since 2004. Labor force participation increased from 60 percent in 2004 to 63 percent in 2007 and employment rates increased from 37 percent in 2004 to 41 percent in 2007. Moreover, unemployment rates decreased from 38 percent in 2004 to 34 percent in 2007.[2] Labor force participation among women increased from 47 percent to 49 percent in the period 2004-2006; whereas such rate has increased from 71 to 75 percent among men in the same period.

Source: Author’s estimates using 2004-2007 LFS data. Note: *2004-2006 are annual data; whereas the 2007 data corresponds to the 3rd quarter 2007. The assessment does not change if we use data for the 3rd quarter for all years.

Furthermore, the economy has managed to create 47 thousand new jobs (in net terms) between 2004 and 2006; especially in the private sector. Private employment has increased relative to non-private employment. While in 2004 private employment accounted for 56 percent of total employment, the proportion increased to 66 percent in 2006. Indeed, the net job creation by the private sector was 80 thousand new jobs between 2004 an 2006. On the other hand, non-private sector employment fell by 33 thousand (net job destruction). The improving situation in regards to job creation continued in 2007. In the third quarter of 2007, the labor market had created 5,700 new jobs as compared to the same quarter in year 2006. Of this increase, 95 percent (about 5,400 jobs) benefited individuals younger than 35 years old, the majority of which were between 15 to 24 years old.

Improvements in the labor market have benefited vulnerable sectors of the population. Improvements in the labor market have benefited low-skilled workers, especially youth, in low-productivity activities. Indeed, Increases in employment and participation rates between years 2004 and 2006 occurred mainly among low-skilled young individuals. Improvements were particularly noticeable among low-skilled youth (15 to 24) engaged in (or entering) unpaid jobs, mainly in the primary sector. By getting unpaid jobs youth acquire experience, which might contribute positively to eventually get a formal job. Most of the improvement in employment and participation rates occurred in regions with a dynamic primary sector, such as Strumika and Bitola. As mentioned before, the employment rate increased from 37 percent to 40 percent between years 2004 and 2006. The largest increase in employment rate (from 13 to 19 percent) occurred among individuals between 15 and 34 years of age having primary or less education. Also, decreases in unemployment rates occurred mainly among individuals between 15 and 34 years of age.

Source: Author’s estimates using 2004-2006 LFS data.

Self employment and unpaid family work have increased their share in total employment, contributing to deterioration in the quality of employment. The composition of employment shows that self employment and unpaid family work have increased as a share of total employment while the share of wage earners to total employment has declined. These changes have contributed to a deterioration of the quality of employment in the country. Indeed, results indicate that wage earners are associated with better employment quality indicators as compared to employers and self-employed. In particular, wage earners are more likely to hold permanent jobs (as proxied by those having an open ended contract), to work full time (40 hours a week or more), and to be formal workers (as proxied by being registered in the pension fund). Furthermore, wage earners display a lower likelihood to be in underemployment (as proxied by the share of workers wanting to work more) and a lower likelihood to have 2 jobs (which is a signal associated to low-quality/low pay in the primary job).

There are large disparities in labor market outcomes across regions. Results of this note indicate a great dispersion of labor market indicators across regions. A large heterogeneity in labor market outcomes within a country is associated with disparities in living standards across regions. In 2006, i) unemployment rates varied from 17 percent in Strumika to 58 percent in Kumanovo; ii) employment rates varied from 28 percent in Kumanovo to 64 percent in Strumika; and iii) participation rates varied from 47 percent in Tetovo to 75 percent in Strumika. Note that Strumika, a traditionally agricultural region, is the one with the better labor market indicators (although employment quality indicators are rather low in the region).

Further analysis may be needed to understand the potential role of internal migration in the regional evolution of labor market outcomes. The lack of convergence across regions points out weakness of equilibrating mechanisms such as wage adjustments and constraints to interregional migration. In principle, workers from worse performing region could migrate to better performing regions. However, there seems to be low internal migration flows in Macedonia. Understanding why this occurs is an area of great interest where further analysis can be developed. Unfortunately, data on internal migration and mobility in the LFS is very limited to conduct such analysis.

Data
This report uses annual data for years 2004, 2005, 2006 and 3rd quarter data for 2007 from the national Labor Force Survey (LFS) conducted by the Macedonian State Statistical Office (SSO). Although data from the 2003 LFS was also available, the existence of methodological differences between this year and more recent ones made it difficult to use these data for trend comparisons. While in the 2003 LFS data were collected once a year in September; from year 2004 and on, data were collected weekly and processed quarterly. Since 2004 the Survey includes a rotational panel, whereby households are interviewed more than one time. The sampling frame, based on the 2002 Census, is stratified, rotational, and “two stage” random. The following Table shows the sample size 15-64 years for the period 2004-2006.
Year / Sample size
15-64 years / Population
15-64 years
2004 / 47,125 / 1,398,504
2005 / 45,219 / 1,411,980
2006 / 42,400 / 1,420,396
Source: Author’s estimates using 2004-2006 LFS data.
There are differences between the World Bank’s and State Statistical Office’s (SSO) estimates of basic labor market indicators. Although basic labor market indicators presented in this article follow a similar time trend as those reported by the State Statistical Office, there are differences in the magnitude. Estimates of labor force participation, employment, and unemployment rates calculated by the State Statistical Office are generally lower in magnitude than those in this article (especially for estimates of participation and employment rates). The discrepancy is mainly driven by differences in the definition of working age population used in the article vs. that used by the SSO. Estimates in this paper follow the ILO definition for working age population (15 to 64 years of age); whereas the State Statistical Office defines working age population as any individual 15 years and older.
LF PARTICIPATION / UNEMP. RATES / EMPLOYMENT RATES
Year / WAP1 / WAP2 / SSO / WAP1 / WAP2 / SSO / WAP1 / WAP2 / SSO
2004 / 59.5 / 52.9 / 52.2 / 38.0 / 37.7 / 37.2 / 36.9 / 33.0 / 32.8
2005 / 61.5 / 54.8 / 54.1 / 38.2 / 37.8 / 37.3 / 38.0 / 34.0 / 33.9
2006 / 62.2 / 55.1 / 55.1 / 36.3 / 36.0 / 36.0 / 39.6 / 35.3 / 35.2
Source: Author’s estimates using 2004-2006 LFS data. WAP1 is defined between 15-64 years; WAP2 is defined 15 years and older.
Data quality: High unemployment rates in FYR Macedonia have often raised questions about the quality of the LFS data. Two reasons that might affect the measurement of the unemployment rate are: (a) non-response that is not properly taken into account in the weights may cause that groups with high/low risk of unemployment are under or over-represented in the sample and (b) a bias in the definition of unemployment. In regards to non-response rates, the European Commission reports a non-response of 17.5 percent for the 2005 LFS. This number is higher than that of other surveys such as Current Population Survey in the US, which has non-response rates of maximum 10 percent. Moreover, the steady reduction in the sample size between years 2004 and 2006 may have contributed to deteriorate the quality of the data and its representativeness. In relation to a possible bias in the definition of unemployment, analysis in this note paid special attention to the way unemployment rates were calculated by the SSO. The definitions used by the SSO were found fully consistent with those used by ILO and Eurostat and no “definition” bias was found. Yet, revising the survey’s sampling may be an important exercise for the SSO since the LFS is the main data source for monitoring the evolution of labor market indicators in FYR Macedonia.

This analytical note is structured in 4 main sections: sections I and II provide respectively an overview of labor force participation and employment in the country from years 2004 to 2006. Section III provides analysis on unemployment trends. Finally, section IV analyses flows between the different statuses of employment analyzed in the previous sections.