gender wage gap in poland: can it be explained by differences in observable characteristics?

22
Karolina Goraus University of Warsaw September 28th, 2012 Gender Wage Gap in Poland – Can It Be Explained by Differences in Observable Characteristics?

Upload: grape

Post on 12-Apr-2017

124 views

Category:

Economy & Finance


1 download

TRANSCRIPT

Karolina GorausUniversity of Warsaw

September 28th, 2012

Gender Wage Gap in Poland – Can It Be

Explained by Differences in Observable

Characteristics?

Outline

� Motivation

� Literature review

� Data and research method

� Raw gender wage gap

� Gender differences in characteristics

� Results of decompositions

� Conclusions

Motivation

� When explaining gender differences in wages, some people may claim that it is due to discrimination, and others that it simply reflects gender differences in some observable characteristics of the individuals that are determinants of wages (Nõpo, 2008)

� To the best of our knowledge there is no empirical research using non-paremetric methods to Poland, while research using parametric methods is scarce

Literature review (1)

� Poland had a significant delay in having their academic, business, and political elites concentrated on the issue of gender differences in the labour market (Grajek, 2003)

� Adamchik and Bedi (2003) doubted if the economic position of females in Poland has improved along with the positive economic performance of the country

� Gender wage gap in Poland over transition : component explained by differences in observed characteristics is quite limited (Grajek, 2003; Adamchik and Bedi, 2003)

Literature review (2)

� Measuring the difference in average wages between males and females is the most basic way to assess gender wage differentials

� Decomposition methods of wage differentialsOaxaca-Blinder decomposition

Two components: one attributable to differences in average characteristics of the individuals, and the other – to differences in rewards that these characteristics have

Literature review (3)

Attempts to refine the Oaxaca-Blinder decomposition

� Male wage structure prevails in the absence if discrimination � Other non-discriminatory wage structures (Neumark, 1988; Oaxaca and Ransom, 1994)

� Outcome variable continuous and unbounded � solution for binary variable (Fairle, 2003), generalization to other discrete and limited variables (Bauer and Sinning, 2008)

� It is only informative about the average unexplained difference in wages � expansion of the method to the case of distributional parameters besides the mean e.g. Juhn, Murphy, and Pierce (1991, 1993), Machado, and Mata (2005), DiNardo, Fortin, and Lemieux (1996), Firpo, Fortin, Lemieux (2007)

Literature review (4)

Attempts to refine the Oaxaca-Blinder decomposition

� Problem: misspecification caused by differences in the supports of the distribution of individual characteristics for females and males

� There are combinations of characteristics for which it is possible to find males but not females in the society, and vice versa. With such distribution of characteristics one cannot compare wages across genders (Rubin, 1977)

� Nõpo (2008) adapted the tool of the program evaluation literature, matching, to construct a non-parametric alternative to Blinder-Oaxaca decomposition method and fix the problem of differences in the supports of distribution of characteristics between females and males

Research method

� Oaxaca-Blinder decomposition��� −��� = �� ̅� − ̅� + (��−��)̅�

∆= ∆� + ∆� +∆� +∆�� Decomposition of Nõpo

� ∆� − can be explained by differences between „matched” and „unmatched” males

� ∆� − can be explained by differences in the distribution of characteristics of males and females over the common support

� ∆� − unexplained part of the gap� ∆� − can be explained by differences between „matched” and „unmatched” females

Data

� About occupational activity of population by demographic and social features

� Comes from the Labor Force Survey performed by Central Statistical Office in Poland and contains quarterly data from 1995q1 to 2011q4

� Persons that are self-employed, unemployed, or inactive, as well as miners and armed forces have been removed from the data set

� Additionally the pooled data set was created� it contains 690414 observations� wages presented in PLN, constant prices of 1995� share of males is 52.5%

� Absolute (PLN, constant prices of 1995) and relative gender wage gap, 1995-2011

� Average hourly wages for females over the years 1995-2011 were 12.5PLN, while for males it was 13.7PLN - the difference amounts to around 9.3 percent of females’ average wage

Raw gender wage gap

Differences in characteristics (1)

Demographic characteristics: Education and marital status

Variables Observations Percent Male Female

Education levels 690 414 100 53 47

Tertiary education 112 697 16 6.5 9.5

High school 82 203 12 3 9

High school vocational 185836 27 13 14

Vocational 240 666 35 24 11

Elementary 69 012 10 6 4

Marital status 690 141 100 53 47

Single 144 305 21 12 9

Married 505 167 73 40 33

Widowed 15 240 2 0 2

Divorced/separated 25 702 4 1 3

Differences in characteristics (2)

Demographic characteristics� Age: females half year older than males� Cities: 40% of females live in the city, while among males the percentage amounts only to 36%

� Mazowieckie: in Mazowieckie region live 10.2% of females and 9.8% of males

Relation of variables to wages:� Age: older people earn more� Cities: people in urban areas earn more� Mazowieckie: people in Mazowieckie region earn more � Education level: highly educated people earn more� Marital status: singles tend to earn less

Differences in characteristics (3)

Job-related characteristics� Occupation

� Very high-skilled occupations (17% of society: higher management, policy makers and specialists): 39% of males, 61% of females

� High-skilled occupations (36% of society: technicians, middle management, office workers, sales and personal services): 33% of males, 67% of females

� Middle-skilled occupations (36% of society: farmers, fishermen, artisans, industrial workers and machine operators): 83% of males, 17% of females

� Low-skilled occupations (11% of society): 43% of males, 57% of females

� Public: 51% of Polish female employees was working in public sector, while for males the percentage was 33%

� Informal: 0.8% of females working in grey economy, while for males the percentage is 1.2%

Differences in characteristics (4)

Job-related characteristics� Branch of economy

� Agriculture (1% of society): 75% of males, 25% of females� Industry (19% of society): 69% of males, 31% of females� Construction (17% of society): 70% of males, 30% of females� Market services (32% of society): 55% of males, 45% of females� Non-market services (31% of society): 31% of males, 69% of females

� Tenure with current employer: 10.7 years for females, and9.8 years for males

� Overall tenure: 17.3 years for females, and 18 years for males� Size of the firm: the same share of females and males in small comapnies, while in medium or large enterprises there is 1.2% more males

Differences in characteristics (5)

Job-related characteristics, relation to wages:� Occupation: people in higher-skill occupations receive higher wage

� Public: higher wages in public sector� Informal: lower wages in informal sector� Branch of economy: highest wages in industry and services� Tenure with current employer/ overall tenure: more tenure results in higher wage

� Size of the firm: higher wages in bigger companies

Intuition: Characteristics of individuals does not seem to explaine gender wage gap in Poland

Results of decompositions (1)

� Decomposition of Nõpo: ∆= ∆� +∆� + ∆� + ∆�

Controls D D0 DM DF DX

Share of

matched

males

Share of

matched

females

Demographic variables 10% 20% 0% 0% -10% 99% 97%

+ Occupation category 10% 20% 0% 0% -10% 96% 93%

+ Industry category 10% 20% -1% -1% -9% 92% 92%

+ Private 10% 21% 0% -1% -10% 99% 95%

+ Informal 10% 21% 0% 0% -10% 99% 97%

+Median tenure 10% 21% 0% -1% -10% 99% 95%

All variables 10% 19% -2% -1% -6% 65% 74%

Results of decomposition (2)

Results of the decomposition (3)

Decomposition of Nõpo based on all variables

Results of

decomposition (4)

Oaxaca-Blinder decomposition

� based on demographic variables : unexplained component of 20%, the same result as in non-parametric

� based on all variables: unexplained component of 21.6%, only slightly sdifferent from non-parametric approach

Results of decompositions (5)

Comparison of decompositions

� Estimators of explained and unexplained gender wage gap in Poland over theperiod 1995-2011 obtained with the use of methodology developed by Nõpohas been confirmed with traditional Oaxaca-Blinder decoposition

� Similar estimators of unexplained component of the gap in Oaxaca-Blinderdecomposition on the whole sample and over the common support

Sensitivity analysis

� Adjusted wage gap bigger than the raw gap� for each wage quartile� for each age category� both in rural and urban areas� in Mazowieckie region and outside� in public and in private sector

� Adjusted wage gap (slightly) smaller than the raw gap� for people with tertiary, vocational and elementary education� for occupations that require more skills� in industry and construction � in informal sector

� Adjusted wage gap within particular groups of society is always positive and vary between 12% and 27%

Conclusions

� Females to a greater extent exhibit characteristics that are well rewarded in the labor market

� Despite better education, they are less frequently employed in better paying positions

� The raw gap over the period 1995-2011 amounts to app. 10%. However, accounting for the differences in endowments the actual wage gap grows to as much as 20%

� Despite covering already 17 years of data, we were not able to identify any clear decreasing trend in gender discrimination in Poland