RRPP Annual Scientific Conference Sarajevo, May 26 th 2012

GENDER PAY GAP IN THE WESTERN BALKAN COUNTRIES: EVIDENCE FROM SERBIA,

MONTENEGRO, AND MACEDONIA

SUNČICA VUJIĆ, MARKO VLADISAVLJEVIĆ,BILJANA APOSTOLOVA, SONJA AVLIJAŠ & NEVENA

IVANOVIĆ

RRPP Annual Scientific ConferenceSarajevo, May 26th 2012

Roadmap of the Presentation About the project…

Motivation, project details, overall objective of the project - what is the main research question and hypotheses to be tested

Literature overview Research methodology GPG facts & figures (EU + 3 WB countries) Results (current and expected) Open questions, suggestions & policy relevance

2

About the Project3

Project Details Organisations:

Foundation for the Advancement of Economics (FREN), Belgrade, Serbia

University American College - Skopje (UACS), School of Business Economics and Management (SBEM).

Project duration: 12 months Starting date: February 2012 Ending date: January 2013

4

Team MembersName Academic Level Position on ProjectMihail Arandarenko Professor, PhD Project CoordinatorMiriam Beblo Professor, PhD Project MentorSunčica Vujić Assistant Professor,

PhDSenior Researcher

Biljana Apostolova PhD candidate Junior ResearcherSonja Avlijaš PhD candidate Junior ResearcherMarko Vladisavljević MA/MSc graduate Junior ResearcherNevena Ivanović MA/MSc graduate Junior Researcher

5

Project Summary This research examines the incidence and determinants

of the gender earnings gap in a sample of Western Balkan Countries (Serbia, Montenegro, and Macedonia).

The gender pay gap (GPG) refers to a difference between the wages earned by women and men.

More formally, GPG is a % difference between average gross hourly earnings of male and female employees, as % of male gross earnings, unadjusted (Eurostat).

6

Motivation 1 Motivation for this study comes from a much more

disadvantageous labour market position of women in comparison to men in countries in the Western Balkans and the necessity to voice the importance of gender equality in the economic arena.

“Women are much less likely to be employed, start a business or advance in the political arena,” (Reva, 2012).

“Women and men are legally equal, but they are not economically equal,” (Apostolova, 2010).

“There is a significant wage gap between men and women in a number of sectors and occupational groups with low educated women being particularly disadvantaged,” (Reva, 2012).

7

Motivation 2 The gender gap in earnings is undoubtedly one of the

most persistent labour market characteristics. In spite of all efforts of the EU to achieve the goal of

equal pay, GPG is about 17.5% on average for the 27 EU Member States (Foubert, 2008; Eurostat, 2010).

GPG in Serbia is between 5% to 12% (lower for higher education levels), in Montenegro about 12.5%, and in Macedonia about 20%.

Further, women have lost out more than men in comparison to the situation before the transition and financial crisis periods, and their position is improving more slowly.

8

Motivation 3 Except for a few recent papers (Reva, 2012; Blunch,

2010; Blunch and Sulla, 2010; Angel-Urdinola, 2008), current evidence in this area is largely descriptive and does not make a systematic comparison between the Western Balkan countries (exceptions are Blunch, 2010 and Apostolova, 2010).

This project aims to fill this research gap. Using Labour Force Survey (LFS), RAD/TRUD surveys,

and own collected data for the three countries, this project aims to shed more light on the observed GPG in Serbia, Montenegro, and Macedonia, and open up relevant policy debates and legislative actions.

9

Economic Development & Social Change

This research falls under the second umbrella topic of the RRPP – Economic Development and Social Change.

What are the benefits of closing the GPG? Redressing the balance on the labour market between

men and women and supporting policies of equal opportunities in the countries of the Western Balkans is important for economic and social development of the countries in question and the whole region. “The economic independence of women is a precondition for

greater influence on decision-making, better access to healthcare, better education and greater protection against potential violence,” (Babović, 2008).

10

Time Plan of the Project 1 Phase 1 (Feb – May 2012)

Literature and existing data review; writing up questionnaires; project website (under construction); administration (Belgrade) and project management (Tirana) trainings; methodology seminar (Skopje); RRPP annual conference (Sarajevo)

Phase 2 (June – August 2012) Half-year report; Survey data collection (IPSOS);

combining different sources of data and analysis; feedback (M&M); workshops 1 (UN Women)

11

Time Plan of the Project 2 Phase 3 (October – January 2013)

Write up academic and policy papers; dissemination of results (check upcoming conference deadlines); workshops 2 (UN Women); possible additional data collection (LFS); feedback (M&M); translation of reports into local languages

Phase 4 (February 2013) End of year report; Final project conference

(Belgrade)

12

Dissemination of Results The findings of this research will be promoted through

website, international conferences, and workshops organised in participating countries, aimed at creating a platform for debate with key stake-holders responsible for labour market policy making and implementation.

Further, results will be distributed in the form of FREN’s working papers, with an aim to publish them in one of the peer-reviewed economic journals (e.g., Economic Journal, Journal of Labor Economics, Industrial and Labor Relations Review, Journal of Public Economics, Journal of Comparative Economics or Economics of Transition).

13

Research Methodology14

Research Methodology Overview Literature review

Hypothesis to be tested Analysis of existing data

Cleaning & merging Collection of own data

Survey design Regression analysis

Mincer earnings regression Oaxaca-Blinder (OB) decomposition

Workshops

15

Literature Overview16

Literature Overview (US, EU, Intl.) GPG and labour market position in the US

Black & Brainerd (2004); Altonji and Blank (1999); Blau (1998); Blau and Kahn (1997; 2000); Loprest (1992)

GPG and labour market position in the EU countries Foubert (2010, EU); Busch & Holst (2009, Germany);

Manning & Saidi (2008, GB); Beblo & Robledo (2008, Germany); Arulampalam et al. (2007, EU); Borbely (2007, BE, NL, HU); Olsen & Walby (2004, GB); Beblo et al. (2003, EU); Machin & Puhani (2003, UK & Germany); Albrecht et al. (2003, Sweden)

International comparisons Weichselbaumer & Winter-Ebmer (2005, Meta analysis); Blau

and Kahn (1992; 1996; 2003)

17

Literature Overview (EE & FSU) GPG and labour market position in EE and FSU:

Rokicka & Ruzik (2010, Poland); Jurajda & Paligorova (2009, Czech Republic); Nestić (2007, Croatia); Jurajda (2003, Czech Republic & Slovakia); Brainerd (2000, EE & FSU); Grajek (2003, Poland); Hunt (2002, East Germany); Orazem and Vodopivec (2000, Estonia & Slovenia)

Legislations and strategies promoting gender equality and equal pay: EC Treaty, Articles 141(1) & 141(3)

18

Literature Overview (WB) 1 GPG and labour market position in the WB countries:

Serbia: Reva (2012); RZS (2011); Blunch & Sulla (2010); Kolin (2010); Kecmanović (2009); USAID (2008); Babović (2008); Arwidsson (2006, Serbia & Montenegro)

Montenegro: IPSOS (2010); MSO (2010) Macedonia: Kazandziska et al. (2012); Angel-Urdinola (2008);

Angel-Urdinola & Macias-Essedin (2008) Country comparisons: Apostolova (2010, Serbia, Macedonia

& Croatia); Blunch (2010, Serbia, EE & CA)

19

Literature Overview (WB) 2 Using data from the UNDP Social Exclusion Survey

(SES), Blunch (2010) compares the gender earnings gap between Serbia and five other countries in Eastern Europe and Central Asia (Kazakhstan, Macedonia, Moldova, Tajikistan and Ukraine). His findings suggest that…

“results are consistent with the presence of earnings discrimination towards females but at the same time also point towards the importance of continued attention towards, among other factors, the education system as a potentially important vehicle for decreasing the gender earnings gap in these former formally gender neutral economies.”

20

Literature Overview (WB) 3 Apostolova (2010) compares gender wage gap between

Serbia, Macedonia, and Croatia based on segregation on the labour market, the employment status, discrimination, and educational differences. She concludes that…

“the relatively low observed gender pay gap in Western Balkan countries may be misleading in judging women’s welfare status in the labour market in at least three aspects. The first is their relatively low employment rate; the second is the pronounced educational advantage of female employees over their male counterparts; and the third is women’s role in terms of family responsibilities.”

21

Definitions, Data & Methodology22

Definition - Unadjusted GPG 1 GPG = % difference between average gross hourly

earnings of male and female employees, as % of male gross earnings, unadjusted (Eurostat). Unadjusted: no correction for national differences in

individual characteristics of employed women and men - no consensus on the method to be used

Hourly wages: to account for differences in hours of work Data sources:

Four-yearly structure of earnings survey (SES) For the intermediate years Member states provide estimates

based on the SES results and annual LFS data. GPG can be calculated on the whole industry level or

different levels of disaggregation (by sector, educational level, age groups, etc.)

23

Definition - Unadjusted GPG 2Detailed methodology of calculation: Industry sectors:

Whole economy except agriculture, fishing, public administration, private households and extra-territorial organisations (NACE Rev 2: B to S excluding O)

Size of enterprises: Only those with 10 employees or more

Type of work: No restrictions on hours worked (part-timers included)

Total earnings:

Include paid overtime and exclude non-regular payments

24

Research Methodology 1 Research involves appending/merging

appropriate waves and variables of the LFS of the three analysed countries in order to: Calculate whole-industry level of unadjusted GPG Calculate GPG for specific groups of men and

women (by level of education, occupation, sector, industry, part/time)

Apply econometric models to answer the proposed research questions.

Prior to 2000s: RAD/TRUD databases

25

Research Methodology 2 The GPG is examined taking into account:

1. level of education; 2. occupation status; 3. part and full time work; 4. industry of work with special focus on agriculture and

contributing family members (mostly women and children);

5. comparing employees, managers, and self-employed (entrepreneurs);

26

Research Methodology 3 The GPG is examined taking into account:

6. formal and informal work; 7. private and public sector;8. comparing periods of pre/after process of privatisation

and restructuring (pre vs. post year 2001); 9. interaction of the afore mentioned transition process

and period of financial crisis which started in 2008; 10. zooming into marginalized social groups (Roma,

refugees, people with disabilities, ex-offenders, etc.)

27

Research Methodology 4 Additional data are collected through the

appropriate survey design (IPSOS). Analysis of the wider labour market factors

that affect GPG: ‘Glass ceiling’ for women and accessibility of jobs

for men and women Possibilities for the advancement on current jobs

for men and women National legislation related to gender equality

28

Methodology - Econometric Models Estimation of male and female earnings equations:

Mincer earnings regression Oaxaca-Blinder (OB) decomposition

29

Methodology - Econometric Models OB three-fold decomposition

Endowments – differences in characteristics (e.g. education levels)

Returns – the return that men and women get (in hourly pay) on the same endowments (e.g. the same educational level)

Interaction – between endowments and returns OB two-fold decomposition

Explained part – group differences in productivity characteristics

Unexplained part – often contributed to discrimination OB overall or detailed (each factor) decomposition

30

Methodology – Workshops FREN + UN WOMEN

Joint regional ‘experts’ workshop, designed to generate dialogue on the methodological and other issues/data limitations, etc.

Workshops with policy/independent bodies such as Ombudsman, gender equality mechanisms, and other stakeholders

Feedback and dissemination

31

Facts & Figures32

GPG 17% EU-27 (Eurostat, 2011)

GPG17%: From 3% in Slovenia to 30% in Estonia

33

GPG 17% EU-27 (Eurostat, 2010)

34

GPG: Facts and Figures (EU) 1 GPG factors (Eurostat, 2010; SES data, 2006)

Low participation of women => low GPG Low GPG is a result of low participation of women

with low wages (low productivity: they would earn lower wages, so they decide to stay home)

High GPG for highly-educated workers (26.1% - tertiary; 13.9% - up to lower secondary) The situation varies among Member States (e.g.

Czech Republic, Greece and France have similar, Belgium and Lithuania have a reverse pattern)

35

GPG: Facts and Figures (EU) 2 GPG factors (Eurostat, 2010; SES data, 2006)

The GPG is lowest for young employees The size of the GPG by economic activities strongly

differs among Member States

36

GPG in Serbia 1 Different results from different surveys: 4.6% (LFS 2009) to

12.4% (UNDP Social Exclusion Survey 2009) Low GPG explained with low participation of women with low skills

(secondary or tertiary education have 60% of employed women and 47% employed men) (Reva, 2012)

Most important determinant of the GPG is education which lowers the unadjusted pay gap (Blunch, 2010)

Different results in different years : 9.2% (2008) vs 4.6% in (2009) Male-dominated sectors have been somewhat more affected by

the global economic crisis

37

GPG in Serbia 2 GPG and education levels: LFS vs administrative data (Reva,

2012 vs SORS, 2010): LFS: GPG is lower for higher education levels (24.0% for primary,

14.5% for upper secondary, 3.4% for tertiary level) Administrative data: GPG is higher for higher educational levels

(15.3% for primary, 12.1% for secondary, 17.7% for tertiary) Age: women have lower wages in all groups except 55-64 (again

due to the effect of low participation of women with high skills) Women are paid much less than men in all occupation groups

and in most sectors (substantial differences in levels) GPG higher in private than in public sector (10.2% and 2.2%),

and in rural areas than in urban areas (12.9% vs 4.5%)

38

GPG in Serbia 3 Oaxaca-Blinder model (Reva, 2012, Blunch, 2010):

Endowments: women have more favourable characteristics (concentrated in better paid jobs, higher educational level) which decreases GPG

Returns: have lower wages for the same characteristics (“returns to endowments”), which increases the GPG

Unexplained part: even larger share of earning differentials; the characteristics of women “hide” the “real” GPG which is higher than unadjusted ones

Detailed decomposition: Most important factor is education, which lowers the gap

through endowments Other observable characteristics widen the gap, but have lower

impact

39

GPG in Serbia 4 Based on the LFS data in 2008 and 2009, Reva (2012)

presents a broad overview of labour market indicators for men and women in Serbia with a focus on employment patterns, entrepreneurship, and career advancement as well as earnings differentials. She finds that…

“wage gap is indicative of discrimination of women in the labour market as earnings differentials cannot be explained by differences in observed characteristics of male and female employees.”

40

GPG in Montenegro “Women receive 87.5% of male hourly wages”

(Ipsos, 2010, LFS data from 2009), GPG is approximately 12.5%

Women receive higher wages only in transporting and storage sectors

Gender is still a significant factor of earnings when education level, age, region, occupation, and industry sector are taken into account (general linear model)

41

GPG in Macedonia 1 42

50-90€ 130-200€ 200-250€ 250-350€ 350-500€ 500-700€0

10

20

30

40

50

60

70

8066.1 63.4 66.9 68.8 67 71

33.9 36.6 33.1 31.2 33 29

Net-Pay and Structure by Gender in % 2009

manwomen

Source: State Statistical Office. LFS (Statistical review: Population and Social Statistics)  

GPG in Macedonia 2 The average net wage paid to men in 2008 and 2009

was higher than the average wage paid to women, with most unequal compensation (gap around 20% and higher).

The wage differences between genders in the country are more pronounced in the private sector, in rural areas and among low educated workers (primary education or less) (Angel-Urdinola and Macias, 2008).

43

GPG in Macedonia 3 The large wage gap between men and women is not

necessarily explained by labour segmentation (whereby women enter sectors offering lower-pay) nor by differences in return to education by gender, but more likely by discrimination.

82.6% of the gender gap in the Macedonia is unexplained (by any observable factor) pointing to a high discrimination against female workers (Angel-Urdinola , 2008).

44

Results & Open Questions45

Results (Current & Expected) Large imbalances between earnings of men and women

when compared across different industries, levels of management, private and public sectors, etc., controlling for socio-economic characteristics.

Results are explained within the framework of labour market milieu (activity, glass ceiling, legislative framework, etc.) of different countries.

However, the discrimination towards women is still highly present in the Western Balkans and its determinants are similar across the analysed countries.

46

Open Questions & Policy Relevance By shedding light on the GPG as one aspect of

economic inequality of women and men, the research will open up debate on other aspects, many of which contribute to GPG, but some of which are the reasons behind the relatively low unadjusted pay gap in the analysed WB countries.

Project will encourage discussion of contributing factors/determinants of GPG as phenomena that can be tackled individually in a targeted manner, through policy/legislative solutions, but also in a coordinated effort with different stakeholders.

47

Open Questions & Policy Relevance In current environments, promotion of soft-policy

initiatives could also be considered by state actors and possibly strengthened by evidence provided by research (such as promotion of business-wide efforts to recruit/promote more women to jobs or higher paying jobs, introduce measures to enable work-life balance for employees, etc).

Legislative action might include re-valuation/assessment of the systematizations of jobs in the public sector and the formulas used to calculate pay levels.

48

Questions?

suggestions?

other?

Sunčica Vujić [email protected] Vladisavljević [email protected] website (under construction)http://fren.org.rs/nuovo/rrgpg/

Thank You!50