Transcript Extensive Literature on Evolution of Gender Wage Gap

Understanding the Evolution
of Gender Wage Gaps
in Ukraine
Ina Ganguli
Harvard University
Katherine Terrell
PREM-Gender, University of Michigan
World Bank Workshop on:
“Women in the ECA Region”
Jan. 24, 2008
Extensive Interest in Evolution of
Gender Wage Gap in ECA

Would the gap grow in transition from
socialist to market-based economy?


Arguments for and against
Evidence is Mixed



Brainerd (2000): grew in 2 FSU but fell in 4
CEEs and no change in 1 CEE
Newell and Reilly (2001): no rise in 16 TEs in
1990s
Orazem and Vodopivec (1995): Fell in Slovenia
Extensive Literature on Evolution
of Gender Wage Gap in ECA

However, changes in gender gaps
are due to many different factors:
Returns to labor
o
Changes in the level of discrimination
(Joliffe, 2002 for Bulgaria; Joliffe and Campos, 2004
for Hungary)
o
Relative changes in returns to HC
(Münich, Svejnar and Terrell, 2005; Liu et al., 2000)
o
Wage-setting policies
(Blau and Kahn,1997 & 2003; DiNardo, Fortin,
Lemieux, 1996)
Extensive Literature on Evolution
of Gender Wage Gap in ECA

Changes in gender gap due to various
factors: Composition of the labor force

Productive Characteristics
(Hunt, 2002; Orazem and Vodopivec, 1995)

Occupational segmentation
(Jurajda, 2003 for CZ; Ogloblin, 1999 for Russia).
Other transition factors, e.g. privatization
(Brainerd, 2002; Liu et al., 2000; Munich, Svejnar and
Terrell, 2005)
Our Research Questions
1. Size of gender gap across the wage
distribution in 1986, 1991 and 2003
2. To what extent are changes in the gaps
due to:
a) Returns (Institutions)
Minimum Wages
Discrimination
b) Composition of labor force
3. Differences in Private v. Public Sector
composition and wage setting practices?
Our Contribution

First micro-economic evidence on Ukraine’s
gender gap during and after communism
o
Look at impact of wage-setting institutions how does the MW affect the gender gap in
Ukraine over time?
o
Previous transition studies focused on the
average gap. We examine the gap across
the distribution.
Ukraine’s Transition
9
8
7
6
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Year
Source: National Bank of Ukraine

Annual Change in GDP (%)
-20
20
-10
30
0
40
10
Minimum Wage as % of Average Wage
-30
10

0

Decline in GDP, hyperinflation,
small change in emp.
Min Wage (‘92)
Entry into the EU
Percent

5
price liberalization; privatizn.

Inflation (Annual Change in Log of CPI, %)
Percent

Independence in 1991
Gradual transition (1992)
Log of CPI, %

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year
Source: UNICEF TransMONEE/IER, Ky iv
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Year
Source: National Bank of Ukraine
Data: Ukrainian Longitudinal
Monitoring Survey (ULMS)




Carried out April-July, 2003
Retrospective questions of jobs in 1986, 1991
Three cross sections (1986, 1991 and 2003) full-time men and
women; over 1,300 men and 1,400 women in each year.
Data issues:

Retrospective:
• Recall error and representativeness of 1986 & 1991
samples

Selection:
• Use of FT workers and people with wage >0

Transition related:
• Inflation
• Wage arrears (12% among men and 9% among
women) but “net contractual monthly salary”
Gender Gap
Public and Private Gaps, 2003


Larger mean gap in public sector; driven by difference
at the top of the distribution
Glass ceiling is most notable in Education, Health &
Social Protection
Three Puzzles
1.
2.
3.
Why did the gap in the lower end of the
distribution fall from the communist period
to the market period?
What explains the persistence of the gaps
at the top end of the distribution (glass
ceiling) from communism to markets?
Why is the a larger gap in the upper end of
the distribution in the public sector than in
the private sector in 2003?
Counterfactual Analysis, Using
Machado and Mata (2004) Method

Method


Create counterfactual densities where women
are given men’s characteristic (Xs) in one
scenario and then women are given men’s
rewards (bs) in another scenario.
Summary of Static Findings

Differences in pay structure (bs) are much more
important than differences in characteristics (Xs)
in explaining the gaps in every year -- explain
more the 75% at each point in distribution
Counterfactual Analysis Over Time:
Summary of Findings

How did changes in the distribution of women’s
Xs change the gaps?
o
o
o

No effect on mean.
Helps reduce gap in the bottom of the distribution
No change in the percentiles at the median and above
Xs at the bottom weren’t as good in 1986, but Xs at
the top were similar  explains puzzles #1 and #2.
How did changes in women’s bs affect change
in the gaps?
o
o
Increased mean gap
However, contributed to a reduction in gap at top and
an increase at the bottom  does not help explain our
puzzles.
Counterfactual Analysis, Over Time:
Summary of Findings

How did changes in the distribution of men’s Xs
over time change the gaps?
o
o

Raised mean gap
 decline in men’s productive characteristics lead to
widening of the gap at bottom 10% but not elsewhere
in the distribution
How did changes in men’s bs affect changes in
the gaps?
o
o
o
Lowered the Gap  Men’s bs declined over time
Contributed to reducing gap in the bottom and
increasing gap at the top
 Helps explain Puzzle #1 - narrowing of the gap at
bottom
Kernel Density Estimates and
Minimum Wages in 1986, 1991, 2003
Counterfactual Analysis, Public vs.
Private: Summary of Findings


In both sectors, gap is mainly due to difference in bs, more
important in Private Sector:
 Private Sector: If women had men’s Bs, mean gap would
have fallen to nearly zero and would have fallen more in
the top half than in the bottom half distribution.
 Public Sector: If women had men’s bs, the mean gap
would have also fallen and more in top half, but effect is
smaller than in private sector.
Differences in Xs small, but composition effect is different in
each sector: If women would have had men’s Xs:
 Private Sector: mean gap would not have changed (but U
shaped across distribn).
 Public Sector: mean gap would not have changed but
would have grown at bottom and fallen slightly at top.
Lower glass ceiling in public explained by women’s
relatively worse characteristics
Another explanation for rise in floor…

Importance of Min Wage for women
Conclusions on Evolution of
Gender Wage Gap in Ukraine



Mean gender gap declined from socially
planned economy (0.40-0.41) to market
driven economy (0.34)
Decline due to narrowing of gap at the
bottom distribution, no change in gap at the
top
Change in structure of LF from public to
private jobs put forces on reducing the gap
at the top
Conclusions on Evolution of
Gender Wage Gap in Ukraine

Explanations:

Decline at bottom due to:
• increase in MWs
• improvement of women’s characteristics
(as those with poor characteristics left the
L.F.)
• decline in men’s rewards

Lack of change at top due to:
• No change in composition of men’s or
women’s characteristics (although bs did
change and contributed to widening)
Conclusions on Evolution of
Gender Wage Gap in Ukraine



In 2003, the public sector had wider mean
gaps than private sector (0.40 v. 0.26)
due to diff. at top, similar gaps at the
bottom
Explanation:



Both sectors rely heavily on MW, especially
for women; MW relatively high in that year
Again, men get much higher rewards for
their labor than women, especially in private
Public sector, women at top of wage
distribution have somewhat poorer
characteristics than men (not in public)