channels of inequality of opportunity: the role of ... of inequality of opportunity: the role of...
TRANSCRIPT
Channels of inequality of opportunity: The role ofeducation and occupation in Europe
Juan Cesar PalominoGustavo Marrero
Juan Gabriel Rodrıguez
Universidad Complutense de MadridUniversidad de La Laguna
34th IARIW ConferenceDresden, August 21-27, 2016
Motivation
What are the channels of inequality of opportunity in Europe?
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 2 / 20
Outline
1 Introduction
2 Methodology
3 Data
4 Results
5 Conclusion
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 3 / 20
IntroductionBackground
Inequality in income due to circumstances considered unfair ((Rawls(1971), Sen (1980), Roemer (1993), Fleurbaey (2008)) → Publicintervention should help to level the playing field (Roemer et al.(2003))
Literature has endeavoured in better measuring IO (Lefranc et al.(2008); Checchi and Peragine (2010); Ferreira and Gignoux (2011);Marrero and Rodrıguez (2012)).
Different approaches (ex-ante and ex-post)Cross-country comparisons
Not much research about the links between circumstances andincome → necessary for public intervention
Insufficient dataNo methodological approach
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 4 / 20
IntroductionContribution
New method to disentangle IO in its principal mediating factors(education and occupation), based on the ex-ante approach and theMLD index, arriving at:
1 = IOREDU + IOR
OCC + IOROTH (1)
Level of education → Up to 30% of IO in some European countries
Occupational category → 1% to 8% of IO depending on the country
Illustrative cross-country correlation analysis between the channelingrole of education and the expansion of higher education
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 5 / 20
MethodologyThe ex-ante IO approach
There is ex-ante equality of opportunity if all types (groups of peoplesharing the same circumstances) face the same set of opportunitiesand have the same mean income: yti = ymi , ∀t,m
Inequality of opportunity is then measured as inequality betweentypes’ means (or parametric estimates)
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 6 / 20
MethodologyDecomposition of the MLD inequality index
The Theil-0 index (Mean Logarithmic Deviation)[IMLD(y) =
1n
∑ni=1 ln(
yyi)] is additively decomposable in between
and within components [Bourguignon (1979); Shorrocks (1980);Foster and Shneyerov (2000); Jenkins (1995)]:
IMLD(y) =
T∑t=1
pt ln(y
yt)︸ ︷︷ ︸
Between→IO
+
T∑t=1
ptIMLD(yt)︸ ︷︷ ︸
Within
(2)
Alternatively, the decomposition can be done parametrically [Ferreiraand Gignoux (2011), Checchi and Peragine (2010)] using thesmoothed distribution µ and the standardized distribution φ:
IMLD(y) = IMLD(µ)︸ ︷︷ ︸Beetween→IO
+ IMLD(φ)︸ ︷︷ ︸Within
(3)
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 7 / 20
MethodologyObtaining the ’smoothed’ distribution
We condition income to circumstances and assume a log-linearrelationship [Ferreira and Gignoux (2011)]:
ln yi = Ciψ + εi (4)
ψ estimates are used to get a smoothed distribution in which allindividuals belonging to the same type have the same expectedincome:
µi = exp[Ciψ] = E(yi | Ci) = yiC (5)
Inequality of this distribution is the inequality of opportunity
IMLD(µi) = IO (6)
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 8 / 20
MethodologyIO channels: Education
Circumstances do not directly convert into future income → theremust exist a set Z of intermediate variables, which are related toindividual circumstances and affect the income of the individual. The’smoothed’ distribution can be expressed as: yCi = f(Zi, νi).
A possible candidate is the level of education Ei, so that Zi = Ei.We can do a parametric estimation of E(yCi | Ei) to measure thatrelation:
ln yCi = Eiη + νi →
{yC,EDUi = exp[Eiη]
yiC,EDU = exp[νi]
(7)
IMLD(yC)︸ ︷︷ ︸
IO
= IMLD(yC,EDU )︸ ︷︷ ︸
IOEDU
+ IMLD(yC,EDU )︸ ︷︷ ︸
IOEDU
(8)
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 9 / 20
MethodologyIO channels: Occupation
yC,EDU is the part of yC not associated with education.
But, is that part of the smoothed income associated with othermediating variables, such as occupation? We can estimate thatassociation with:
ln yC,EDUi = Oiκ+ ξi →
{yC,OCCi = exp[Oiκ]
yiC,OCC = exp[ξi]
= yiC,OTH (9)
IMLD(yC,EDU )︸ ︷︷ ︸
IOEDU
= IMLD(yC,OCC)︸ ︷︷ ︸
IOOCC
+ IMLD(yC,OCC)︸ ︷︷ ︸
IOOTH
(10)
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 10 / 20
MethodologyDisentangling IO
Decomposing the smoothed distribution.
µi = Ciψi︸︷︷︸ln yCi
= Eiη︸︷︷︸ln yC,EDU
i
+ νi︸︷︷︸Oiκ︸︷︷︸
ln yC,OCCi
+ ξi︸︷︷︸ln y
C,OTHi
(11)
Decomposing IO
IMLD(yCi )︸ ︷︷ ︸
IO
= IMLD(yC,EDUi )︸ ︷︷ ︸
IOEDU
+ IMLD(yC,OCCi )︸ ︷︷ ︸
IOOCC
+ IMLD(yC,OTHi )︸ ︷︷ ︸
IOOTH
(12)
Obtaining the IO shares (dividing byIMLD(yCi ) = IO):
1 = IOREDU + IOR
OCC + IOROTH (13)
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 11 / 20
DataDatabase
EU-SILC framework
2005 and 2011 intergenerational transmission of advantage modules(data from 2004 and 2010)26 european countries in both wavesAdvantage: equivalent household income (household heads aged 30-50)
Circumstances
Parental education (4 levels; father and mother), parental occupation(11 one-digit categories; father), gender, inmmigration status (national,other-EU, non-EU), financial difficulties percieved during household (5levels).
Channels of transmission
Education (6 ISCED levels)Occupation (11 ISCO-88 categories)
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 12 / 20
ResultsInequality of opportunity
Nordic countriesat the bottom(except Iceland2010)
EL, LU and PTat the top (2010)
No clear patternovertime (EL,BE, IS, SKincrease; PT, PL,LT and LVdecrease)
In general, in linewith Marrero andRodrıguez (2012)and Brzezinski(2015)
AT
BE
CY
CZ
DE
DK
EE
EL
ES
FI
FR
HUIE
IS
IT LT
LU
LV
NL
NO
PL
PT
SE
SISK
UK
0.00
0.01
0.02
0.03
0.04
0.05
0.00 0.01 0.02 0.03 0.04 0.05Inequality of Opportunity 2004
Ineq
ualit
y of
Opp
ortu
nity
201
0
Region
Central
Eastern
Mediterranean
Nordic
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 13 / 20
ResultsThe educational channel
Between 8 and30% of IO
PT and LU at thetop
Nordic countriesat the bottom(except NO)
No clear overtimetrend (EL, LU,DE, NO increase;SE, FI, ISdecrease)
AT
BECY
CZ
DE
DK
EE
EL
ES
FI
FR
HU
IE
IS
IT
LT
LU
LV
NL
NO
PL
PT
SE
SISK
UK
0%
10%
20%
30%
40%
0% 10% 20% 30% 40%Share of IO transmitted by Education 2004
Sha
re o
f IO
tran
smitt
ed b
y E
duca
tion
2010
Region
Central
Eastern
Mediterranean
Nordic
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 14 / 20
ResultsThe occupational channel
Between 1 and8% of IO (1-5%in mostcountries)
CY and AT atthe top
No geographicalpattern
No clear overtimetrend (CY, ATincrease; NO,DE, HU, IEdecrease)
AT
BE
CY
CZ
DE
DK
EE
EL
ES
FIFR
HU
IEIS
ITLT
LU
LVNL
NO
PL
PT
SE
SI
SK
UK
0.0%
2.5%
5.0%
7.5%
0.0% 2.5% 5.0% 7.5%Share of IO transmitted by Occupation 2004
Sha
re o
f IO
tran
smitt
ed b
y O
ccup
atio
n 20
10
Region
Central
Eastern
Mediterranean
Nordic
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 15 / 20
ResultsThe educational channel and the extension of highereducation (I)
High negativecorrelationbetween theshare of thepopulation withhigher educationand the share ofIO channeledthrougheducation
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
AT BE
CY
CZ
DEDKEE
EL
ES
FI
FR
HU
IE
IS
IT
LT
LU
LVNL
NO
PL
PT
SE
SI
SK
UK
AT
BECY
CZ
DEDK
EE
EL
ES
FI
FR
HU
IE
IS
IT
LT
LU
LV
NL NO
PL
PT
SE
SISK
UK
y = 0.27 − 0.42 ⋅ x, r2 = 0.342
0%
10%
20%
30%
40%
0% 20%
40%
60%
Share of Sample with High Education
Sha
re o
f IO
tran
smitt
ed th
roug
h E
duca
tion
Wave
●
●
2004
2010
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 16 / 20
ResultsThe educational channel and the extension of highereducation (II)
High positivecorrelationbetween theshare of thepopulation withlower educationand the share ofIO channeledthrougheducation
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
AT
BE
CY
CZ
DEDKEE
EL
ES
FI
FR
HU
IE
IS
IT
LT
LU
LV NL
NO
PL
PT
SE
SI
SKUK
ATBE
CY
CZ
DE
DK
EE
EL
ES
FI
FR
HU
IE
IS
IT
LT
LU
LV
NLNO
PL
PT
SE
SI
SK UK
y = 0.075 + 0.36 ⋅ x, r2 = 0.475
0%
10%
20%
30%
40%
0% 20%
40%
60%
Share of Sample with Low Education
Sha
re o
f IO
tran
smitt
ed th
roug
h E
duca
tion
Wave
●
●
2004
2010
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 17 / 20
Concluding remarks
A new strategy to disentangle IO channels
Importance of the educational and occupational channels
The educational channel and the expansion of education
Prospective research:
What could explain the remaining share of IO?
Apply this strategy to new mediators and other measures of IO whendata is available
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 18 / 20
Thanks!
Juan C. Palomino([email protected])
Gustavo A. Marrero([email protected])
Juan G. Rodrıguez([email protected])
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 19 / 20
Channels of inequality of opportunity: The role ofeducation and occupation in Europe
Juan Cesar PalominoGustavo Marrero
Juan Gabriel Rodrıguez
Universidad Complutense de MadridUniversidad de La Laguna
34th IARIW ConferenceDresden, August 21-27, 2016
References I
Bourguignon, F. (1979). Decomposable income inequality measures.Econometrica: Journal of the Econometric Society , 901–920.
Brzezinski, M. (2015). Inequality of opportunity in Europe before and after theGreat Recession. University of Warsaw - Working Papers No.2/2015 (150).
Checchi, D. and V. Peragine (2010). Inequality of opportunity in Italy. TheJournal of Economic Inequality 8(4), 429–450.
Ferreira, F. H. and J. Gignoux (2011). The measurement of inequality ofopportunity: Theory and an application to Latin America. Review of Incomeand Wealth 57(4), 622–657.
Fleurbaey, M. (2008). Fairness, responsibility, and welfare. Oxford UniversityPress.
Foster, J. E. and A. A. Shneyerov (2000). Path independent inequality measures.Journal of Economic Theory 91(2), 199–222.
Jenkins, S. P. (1995). Accounting for inequality trends: decomposition analysesfor the UK, 1971-86. Economica, 29–63.
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 1 / 2
References II
Lefranc, A., N. Pistolesi, and A. Trannoy (2008). Inequality of opportunities vs.inequality of outcomes: Are western societies all alike? Review of Income andWealth 54(4), 513–546.
Marrero, G. A. and J. G. Rodrıguez (2012). Inequality of opportunity in Europe.Review of Income and Wealth 58(4), 597–621.
Rawls, J. (1971). A Theory of Justice. Cambridge, Ma: Harvard University Press.
Roemer, J. E. (1993). A pragmatic approach to responsibility for the egalitarianplanner. Philosophy Public Affairs (20), 146–166.
Roemer, J. E., R. Aaberge, U. Colombino, J. Fritzell, S. P. Jenkins, A. Lefranc,I. Marx, M. Page, E. Pommer, J. Ruiz-Castillo, et al. (2003). To what extentdo fiscal regimes equalize opportunities for income acquisition among citizens?Journal of Public Economics 87(3), 539–565.
Sen, A. (1980). Equality of what?, Volume 1. Tanner Lectures on Human Values.
Shorrocks, A. F. (1980). The class of additively decomposable inequalitymeasures. Econometrica: Journal of the Econometric Society , 613–625.
Palomino/Marrero/Rodrıguez (UCM/ULL) Channels of inequality of opportunity IARIW 2016 Dresden 2 / 2