You are on page 1of 13

1

THE QUESTION OF THE


GENDER WAGE GAP IN ITALY
A STUDY ON THE DATA OF THE BANK OF ITALY

PIERO DELLA GIUSTINA


MASTER IN ECONOMICS

University of Trento – AY 2013/2014


Gender Gap around the World
2

World average Ranking

(1) Iceland
0,97
Health (2) Finland
0,96
...

(5) Philippines
0,92
Education ...
0,93
(30) Spain

...
0,15
Politics (71) Italy
0,21
...

(105) Japan
0,56
Economy ...
0,6
(136) Yemen

0 0,2 0,4 0,6 0,8 1 1,2 0 0,2 0,4 0,6 0,8 1 1,2

2006 2013 Health Education Politics Economy

Source: World Economic Forum, The Global Gender Gap Report 2013
Unadjusted Gender Pay Gap in Europe
3

35
Unadjusted GPG:
30
Mean male log hourly
salary minus mean 25
female log hourly
salary: 20
17,3 17,2
16,5
wm − w f
16,2 16,4

15

Does not take into


10
account differences in 6,7
the endowment of 5,1 4,9
5,5 5,3 5,8
5
market-rewarded
characteristics
0
(education, age, expe 2007 2008 2009 2010 2011 2012
rience etc.)
EU (27) Czech Republic Estonia Italy Austria Poland Portugal

Source: SES (Eurostat)


The European Labor Market
4

Female participation rate (20-64) Female part-timeover the total employment (15-64)
85
Netherlands

80 Switzerland

EU(28) 2013
75
75,8
Italy 2008
72,8
70 2004
69,8 Finland

65 62,8 Slovakia
62,6
59,4 0,0 20,0 40,0 60,0 80,0
60
Weekly hours in household work by income
55 provision (2010)
50,5 49,9
50 Female 100%
50,6
48,3
45 Female >60%

Equal (40-60%) Women


40
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Men
Male >60%

Iceland Norway EU (28 countries) Male 100%


Italy (men) Italy Greece
0 10 20 30 40 50

Source: LFS (Eurostat) and EU-SILC


Estimation Techniques
5

Pooled OLS estimation:


wage equation à la Mincer with dummy for gender. Measures the wage penalty/premium associated with the
observation being a woman/man ceteris paribus:
w i = X iT β + α Fi + u i

Oaxaca-Blinder decomposition:
splits the unadjusted gap in endowment effect and coefficient effect (adjusted gap).
(1) separate estimation of wage equations by gender:
w i = X ig ,T β + ui
g g g

g = m, f
(2) decomposition of the unadjusted gap:
m f
(
m f
w − w = X − X βˆ m + X βˆ m − βˆ
f
) ( f
)
Heckman correction:
corrects for the non-randomness of the sample of women in employment:
( ) (
E w i | X iT , f , Pi = 1 = X iT , f β g + E u i | Pi = 1
f f
)
To ensure the second term is 0, we include an additional regressor in the female wage equation (Inverse Mill’s ratio):
w i = X i f ,T β + θλ i + u i
f f f

Mincer (1974), Oaxaca (1973), Blinder (1973), Heckman (1976, 1979)


Italian literature
6

Reference Data Ref. Model Unadj. Gap Adj. Gap Endowm.


source Year
Addis & SHIW 1989 Pooled OLS estimation on annual gross 12.1
Waldman income
(1996)

Pissarides et ECHP 1998 Oax-Bli on hourly gross wage (with 8.5 15.9
al. (2005) selection correction)

Addabbo & ECHP 2007 Oax-Bli on hourly gross wage (with 5.5 14.5 -9.0
Favaro selection correction)
(2007)

Centra & ISFOL 2007 Oax-Bli on hourly net wage (with 8.7 10.8 -2.1
Cutillo selection correction)
(2009)

Zizza SHIW 1995- Pooled OLS estimation on hourly net 4.9 to 7.7 8.7 to 12.3
(2013) 2008 wage (with selection correction)
SHIW database 2012
7

I analyze a sample of men and women either in dependent employment or housewives from the 2012 wave
of the Survey of Household Income and Wealth (Bank of Italy). The sample is made of:

Dependent workers Hosewives Avg hourly log wage


Male 2787 0 2.195
Female 2238 1674 2.144
Total 5025 1674

The survey makes available a number of controls to be included in the estimation of the participation and wage
equations:
Gender distribution by educational achievement Gender distribution by employment contract
Doctorate 100%
Master 90%
Bachelor 80%
Highschool 70%
Vocational High. 60%
Middle 50% female
Elementary 40% male
No edu. 30%
20%
0 0,1 0,2 0,3 0,4 0,5 10%
0%
Female (not in emp.) Female (in emp.) Male
part-time public sector
SHIW database 2012
8

Distribution genders by occupation (% of Distribution female pop. in/out emp. by area


tot. population in employment) 0,4
0,35
0,6
0,3
0,25
0,2
0,5 0,15
0,1
0,05
0,4 0
north-west north-east center south islands

not in emp. in emp.


0,3
Distribution in/out emp. by marital status
100%
90%
0,2 80%
70%
60%
50%
0,1 40%
30%
20%
0 10%
0%
blue collar clerk teacher manager executive
married not married divorced not divorced
male female
not in emp. in emp.
Econometric model and results
9

Oaxaca-Blinder decomposition with Heckman correction:

w
m
−w
f
= X ( m
− X
f
)βˆ m
+ X
f
(βˆ m
− βˆ f
)− θ λ
)

Results:
Log hourly wage (men) 2.195
Log hourly wage (women) 2.144
Undajusted gap 5.11%
Adjusted gap (coeff. effect) 8.39 to 9.47%
Endowment effect -2.11 to -3.28%
Correction term -0.82 to -1.36%
Segregation 0.49 to 2.09%
Results (selected variables)
10

-0,11 -0,09 -0,07 -0,05 -0,03 -0,01 0,01 0,03 0,05 0,07 0,09 0,11 0,13 0,15

age 35to44
In favor of women In favor of men age 45to54
age 55to64
n workexp
exp
exp^2
citizen
edu elem
edu middle
edu high_voc
edu high
edu bach
edu master
edu post
mu med
mu big
mu hu
part time
permanent
fixed term
public sec
PF
constant

Endowment Returns
Critiques
11

Measurement error
Unobserved heterogeneity (ability, risk-preferences
etc.)
Actual experience
Correction applied
Simultaneous causality (human capital investments)
Controls included
Segregation and level of aggregation
Endogeneity of the family choices
Coelli (2014)
Consequences
12

Pension gap (2011)


Psychological: dissatisfaction, less
0 10 20 30 40 50
sense of self-fulfilment
Luxembrug
Personal/family decisions: Germany
EU(28)
underinvestment in human Italy
capital, child and family-care Denmark
Estonia
Economic (short term): less
availability of economic resources GDP growth with participation gap reduction
2 700
(hourly wage, annual income
etc.), less economic independence 2 400

Economic (long term): gap in


pensions 2 100

Economic (large scale): foregone 1 800

economic growth, ageing labour


force 1 500
2011 2016 2021 2026

No-change reduce by 50% reduce by 100%

Source: EU-SILC and OECD


13 Thank you for your attention!

You might also like