Paper Income Inequality Dynamics Since NAFTA.pdf
-
Upload
gerardo-esquivel -
Category
Documents
-
view
216 -
download
0
Transcript of Paper Income Inequality Dynamics Since NAFTA.pdf
-
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
1/35
The Dynamics of Income Inequality in Mexico since NAFTA [with Comment]Author(s): GERARDO ESQUIVEL and Guillermo CrucesSource: Economa, Vol. 12, No. 1 (Fall 2011), pp. 155-188Published by: Brookings Institution PressStable URL: http://www.jstor.org/stable/41302974.
Accessed: 14/02/2014 15:01
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at.http://www.jstor.org/page/info/about/policies/terms.jsp
.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
.
Brookings Institution Pressis collaborating with JSTOR to digitize, preserve and extend access toEconoma.
http://www.jstor.org
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/action/showPublisher?publisherCode=brookingshttp://www.jstor.org/stable/41302974?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/41302974?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=brookings -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
2/35
GERARDO
ESQUIVEL
The
Dynamics
of
Income
Inequality
in
Mexico since NAFTA
That
at
nineteenth
least
Mexico
since
century
is a
Alexander
highly
that
unequal
the
von
region
Humboldt
country
then
is
known
wrote
a fact hat
as
at the
New
has
beginning
been
Spain
recognized
was
of
"the
the
t least since Alexandervon
Humboldtwrote t the
beginning
f the
nineteenth
entury
hat he
region
thenknown
as New
Spain
was "the
country
f
nequality." adly,
this s still rue
n
the
wenty-firstentury.
or
example,
Corbacho and Schwartz
2002)
point
out that Mexico's income
inequality
s
significantly
ore
pronounced
han heLatinAmerican
verage,
which s theregionwith hehighest egreeof nequalityntheworld."Also,
looking
at the
ong-run
rend n income
inequality
n
Mexico leaves little
room to be
optimistic.
istoricaldata show thatwhile Mexico achieved an
important
eduction
n
nequality uring
he
1960s
and
1970s,
periods
f
very
rapid
economic
growth,
he
country
as
experienced ery
ittle
progress
n
incomedistributionince the 1980s
(Sz6kely
2005).
This
situation, owever,
may
have started o
change
n recent
years.
This
paperprovides
videnceon thereductionn
ncome
nequality
hathas taken
place
in
Mexico since 1994 and discusses some of the
ikely
ourcesof this
trendsee figure ),which s importantor t easttwo reasons:first,ecause
it has almost
ompletely
eversed he
widely
documentedncrease n
nequal-
ity
hat ccurred
n
the1984-94
period
Bouill6n,
Legovini,
nd
Lustig
2003;
Legovini,
ouill6n,
nd
Lustig
005),
and
second,
ecause the eductioneems
to be theresult
f two
mportant
tructural
hanges
n the
Mexican
economy:
Gerardo
squivel
s with l
Colegio
eMexico.
This
aper
was
originally
ritten
s
part
ftheUnited
ations
evelopmentrogram
(UNDP)
roject
Markets,
he
tate,
nd he
ynamics
f
nequality,"
oordinated
y
Nora
Lustig
nd uis
elipe 6pez-Calva.
thank
articipants
n
previous
eminarst
UNoffices
n
New ork ndMexico ity,swell sparticipantst he atin mericannd aribbeanco-
nomic
ssociation
LACEA)
ongresses
n
Riode Janeiro
ndBuenos ires or heirseful
comments
nd
uggestions.
am
articularly
hankfulo
GuillermorucesndAdriana
ugler
for heirommentsnd
ecommendationsn
previous
ersionf his
ork.dith ortes
ro-
vided
ruly
utstanding
esearch
ssistance.he sual
isclaimers
pply.
155
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
3/35
156
ECONOMIA,
all 011
FIGURE
1. Mexico'sin oefficientnd heilndex
Source:EMASnd orfdank
2010).
the rrival
f
better-targeted
ocial
programs
uch s
Progresa/Oportunidades
and a reductionn abor ncome nd
wage inequality
hat eems to be associ-
atedwith he
mprovement
n educational evels n Mexico. A third ontribut-
ing
factor o therecent eduction
n
inequality
as been the
growing
low f
remittanceshat
many
Mexicans
iving
broad end o their amilieseft ehind
inMexico.
The
possibility
hat oth ocial
policy
nd educational
mprovementsar-
tially
xplain
thereductionn income
nequality
n Mexico cannotbe under-
estimated.n
fact,
ncome
nequality
s
diminishing
n several atinAmerican
countries,
nd it s
possible
that imilar actors ould be at
play
in
many
of
those ountries.1 his could ead not
only
to an
appropriate
valuation
f the
new social
policies
that re
being mplemented
n
the
region
but also
to
a
reconsiderationf the effect hat
higher
evels
of
education,
ombinedwith
1.
See,
or
xample,
erreira,eite,
ndWai-Poi
2007)
nd anos nd thers
2010)
or
the
razilian
ase;
Gasparini
nd ruces
2010)
or
rgentina;
nd berhaidnd
ngel
2008)
for hile.
or
more
eneral
iew n ecentrendsn
nequality
n atin
merica,
ee hentro-
ductiono
Ldpez-Calva
nd
ustig
2010)
s well s
Gasparini,
ruces,
nd ornarolli
2009).
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
4/35
Gerardo
squivel
157
TABLE 1. OverviewfMexican
conomy
Stage Stage
I
Stage
II
Stage
V
Stage
Feature
1950-70 1970-82 1982-94
1994-2000
2000-06
Macro
High
rowth Highrowth
Low
rowth
1994risisnd Low
rowth
withacro-
with acro-
with acro-
recovery;
ow withacro-
economic
economic
economic
growth
ith
economic
stability instability
adjustment
somenflation
stability
Openness
Semidosed Semidosed Unilateral
NAFTA;
reerade
Open
conomy
economy
ith
economy
ith
openness
agreement
tariffnd
tariffnd in
985;
withhe U
nontariff nontariff beginning
barriers barriers of AFTA
negotiations
in 992
Inequality
Relatively
table
Rapid
eduction ncrease
Reduction Reduction
Social
Low
Low
Nontargeted
argeted Expansion
f
protection
social
programs
n
Progresato
programs:
ruralreas:
urban
Solidaridad
Progresa
nd areas
Procampo
Source:
uthor's
ompilation.
globalization
nd
trade
iberalization,
ay
have on
nequality
n
middle-income
countries.2
The
paper
first
rovides
an
overview of macroeconomic onditions n
Mexico
during
he
past
decades followed
by
estimates
f ncome
nequality
in
Mexico
using
alternative efinitionsf income.
Next,
a Gini
decomposi-
tion
nalysis
s conducted
o
nvestigate
hecontributionf differentncome
sources o
the volution f
nequality
n
Mexico,
and therole
of
ncome
abor
andwage inequalitys discussed n explaining hedynamics f inequality.
The
conclusions ollow.
An
Overview f Mexico'sEconomic onditions ince 1950
Table 1
provides
n
oversimplifiedummary
f Mexico's economic
perfor-
mance
ince
1950.
n
thefirst
tage
1950-70),
GDP
grew
t a
relatively apid
pace
(3
percent yearpercapita),
with
price tability,
ow fiscal
deficits,
nd
a fixedexchangeratesince 1956. The second stage (1970-82) was again
2. See
Goldberg
nd
avcnik
2007)
or recent
urvey
n his
ssue.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
5/35
158 ECONOMI
,
Fall 011
a
period
of
rapid
growth
3
percent
year
per capita),
but withmacro-
economic
instability.
uring
this
period,
Mexico
suffered
ouble-digit
annual
nflation
nd
arge
devaluations
n
both1976
and 1981.
Mexico's
gov-
ernment
ncurred
arge
fiscal
deficits,
nd
public
sector
xternal ebt
oared.
The
two initial
tages
were characterized
y
a
semi-closed
conomy
with
high
ariff
nd nontariffarriers.
uring
he first
tage, nequality
emained
relatively
table,
whereas
during
he
econd
stage
here
was a
rapid
reduction
in
ncome
nequality
n Mexico
(Szekely
2005).
The third
tage 1982-94)
was
one of structural
djustment
nd
mportant
economic reforms.
uring
this
period
Mexico
went
through process
of
macroeconomic
djustment
hat ed to a radical
change
in its economic
model:
the
government
rastically
educed
publicexpenditures;
herewas
an
important
enegotiation
f
public
sector
foreign
ebt;
arge-scaleprivatiza-
tiontook
place;
and,
n
1985,
in
the midst
f an
unexpected
ollapse
in the
price
of ts main
export
oil),
Mexico
unilaterally pened
up
its
economy
y
significantly
educing
tstariffs
nd
eliminating
most f ts nontariff
arriers.
In
the
early
1990s,
Mexico
announced ts
intention f
going
well
beyond
those eformsand ocking hemn)byproposing
free rade
greement
ith
the
United
tates
nd Canada.3
The
agreement
ent nto
ffect
n
1994 as
the
North
AmericanFree
Trade
Agreement
NAFTA),
establishing
he
argest
free rade rea
in theworld and the
most
symmetrical.
uring
his
period
the
Mexican
economy
tagnated
n
per capita
terms
nd income
nequality
increased
ubstantiallyhroughout
he
period.
The first
ear
of the
post-NAFTA
period
was characterized
y
a severe
macroeconomic
risis
that
egan
n
December 1994.
In
that
month,
Mexico
experienced
large
devaluation
nd was close
to
ncurring
financial
efault.
The fiscal ndmacroeconomicdjustmentf 1995 led to a sharp nd steep
decline
n
economic
activity uring
995
(a
contraction
f 8
percent
n
per
capita
GDP).
Later,
rom
995
to
2000,
thedomestic
conomy
ecovered
el-
atively uickly,
mainly
ecause of
an
important
ncrease
n Mexican
exports
to
the
U.S.
market.
etween
1995
and
2000,
Mexico's
per
capita
GDP
grew
at a rate
of
4
percent
year.
The
first
ost-NAFTA
stage
was also
characterized
y
the
mplementa-
tionof two
mportant
ocial
and economic
programs:
rogresa
later
known
as
Oportunidades)
nd
Procampo.
The first
rogram
s a
targeted
onditional
cash transferrogram hat tartedn 1997 and is currentlyonsidered he
most
mportant
ntipoverty
rogram
n Mexico.
Progresa
was
implemented
3.
SeeTornellnd
squivel
1997)
ormore etailsn hese
ssues.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
6/35
Gerardo
squivel
159
first
n
rural
reas,
although
t has includedurban reas since 2001. The sec-
ond
program,
rocampo,
s an
income-support rogram
or
gricultural ro-
ducers
designed
o
help
them
ace thetransition rom closed
economy
o a
more
open
economy.
The
program,
which
began
in
1994
when NAFTA
went
nto
effect,
s considered
badly designed program
n
redistributive
terms
Esquivel, Lustig,
nd
Scott
2010).
On
average,
he
period
1994-2000
was one of mediocre conomic
performance
2
percent rowth
year),
but
it was also the
period during
which ncome
inequality
tarted o fall. The
mostrecent tage,from
000 to
2006,
was one
of low
growth
withmacro-
economic
stability. uring
those
years,
Mexico's
per capita
GDP
grew
at
only
1
percent
year,
ecause
twas
negatively
ffected
y
theU.S. recession
of 2000-01.
Nevertheless,
uring
his
period
ncome
nequality
was reduced
even further.
Income
nequality
n
Mexico
ince
NAFTA
Before nequalitynMexico is discussed, t s importantoclarifywhatmea-
sure f
nequality
nd what
efinitionf
ncome s used
n
this
aper,
incedif-
ferent efinitions
ould lead
not
only
to different
stimates f
nequality
ut
also to
slightly
ifferentonclusions.4
ost of
the
results,
owever,
re
robust
to alternative efinitions
f ncome
nd alternative
easures f
nequality.
In
this
paper
the Gini coefficient
s used as
the
preferred
measure
of
inequality.5
his measure
not
only
satisfies
ll the desirable
properties
f an
inequality
measure6 ut
also can be
decomposed
by
ncome
ource,
feature
that
s of nterest ere.On
theother
and,
nequality
s
usually
measured
sing
either urrentotal ncomeor currentmonetaryncome.7 oth definitionsre
used
n
the nitial stimates
f
nequality,
ut ater
monetary
ncome stimates
are the
only
focusof attention.
igure
2
provides
simpledescription
f the
4.
Corbacho
nd chwartz
2002)
nclude
survey
f
Gini oefficientstimates
nMexico
for ifferent
eriods
nd ifferentncomeefinitions.
5. TheGini
oefficientoes
not
apture
ell
hanges
t he xtremes
f he istribution.
However,
ther easures
f
nequality
hathow rends
imilaro hose escribed
n he ext
are vailableromhe
uthorn
equest.
ee
also
Campos
2008)
or
omparisons
sing
lter-
native easuresf
nequality.
6. These rinciplesre s follows:dherenceo he igou-Daltonransferrinciple;ym-
metry;ndependence
f
cale;
omogeneity;
nd
ecomposability.
7.
Theres thirdefinitionf ncome
hats
widely
sed
nMexico: et otalncome.
his
definitions similaro
urrentotalncomeut educts
ifts
nd n-kindransfers.
hismea-
sures the ne sed
n
he
fficialstimationf
poverty
atesnMexico.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
7/35
160
EC0N0MIA,
Fall 011
FIGURE 2. Sourcesf urrentotalncomenMexico
f
^
-
Laborncome
-
Own usinesses
Monetary
Income
-
Assetsncome
*
Remittances
Current .
publicransfers
Total
J
-Transfers
Income '
^
pens,ons
rivateransfers
Non-
Monetary
ncome
V
Source:NIGH
various
ears).
components
f both ncome
definitions.
he
description
f the sources of
monetary
ncome are laterused in a
Gini
decomposition
xercise.All esti-
matesuse information
rom he National
Survey
of Household ncome and
Expenditure
ENIGH,
the
cronym
n
Spanish). Surveys
re available for he
years
1984, 1989, 1994,
1996, 1998, 2000, 2002, 2004, 2005,
and 2006.
Figure
3 shows the
evolutionof the Gini coefficientn Mexico forthe
period
1984-2006,
using
alternative efinitions f income. The
figure
clearly
hows the
existence
of
an inverted
shape
that
eaks
in
1994
in all
cases
and that
teadily
declines thereafter.he
figure
lso
shows the
rapid
increase n
inequality
hat ook
place
between 1984 and
1994,
which has
been
reported
n,
among
other
studies, Bouill6n,
Legovini,
and
Lustig
(2003) andLegovini,Bouill6n,andLustig 2005). The Gini coefficient or
current
monetary
ncome
dropped
from .564 in 1994 to 0.505 in
2006,
a
10
percent
eduction;
he
corresponding
measurefor otal ncome
dropped
from
.537 to
0.494,
an 8
percent
eduction. hese
reductions re similar n
magnitude
o those
recently
bserved
n
Brazil and documented n Barros
and others
2010).
In
annual
terms,
nequality
n Mexico has fallen t a rate
of 0.9 and 0.7
percent year
n
the case of current
monetary
ncome and
total
ncome,
respectively. lthough
hose rates are stillbelow the annual
1
percent
ate at which ncome
nequality
diminished
n
Mexico between
1954 and1984,they how significantmprovementith espect o the 1994
figures.
Figure
3 also
shows
a
few other
nteresting
esults. or
example,
he
fig-
ure
shows that hedistributionf
monetary
ncome
s more
unequal
that
s,
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
8/35
Gerardo
squivel
161
FIGURE 3 Mexico:ini oefficientsor lternativencome
efinitions,
984-2006
Source:uthorslaborationasednNIGHvariousears).
it has a
higher
Gini
coefficient)
han
hedistributionf total urrentncome
(which
s
explained y
thefact hat
onmonetary
ncome s less
unequally
is-
tributed);
n the ther
and,
he
figure
lso shows hat efore ither ransfers
r
remittancesre
ncluded,
heGini coefficientf
monetary
ncome
ends o be
slightly igher
han heGini coefficientf
monetary
ncome,
hus
uggesting
the
finalizing
ontribution
f these wo
factors,
n issue discussed ater
n
the
paper.
The
Urban/Rural
imensionf
nequality
n
Mexico
Previous tudieshave
shown herelevance f
understanding
he
dynamics
f
rural
ndurban
nequality
n Mexico
separately.
or
example,
nuco-Laguette
and
Szekely
(1996)
showed that
nequality
withinurban and rural
areas
accounted or84
percent
f total
nequality
n Mexico
in
1992,
whereas
nly
one-sixth ftotal
nequality
as
explained
y
the ural/urban
ap.
For that ea-
son,thispapernow focuseson thedynamics f nequalityn rural nd urban
areas in Mexico since 1994. As discussed
ater,
his
distinction
s crucialto
understanding
he contributionf differentactors
n
the recentdownward
trendn
nequality
n the
ountry.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
9/35
162
ECONOMIA,
all 011
FIGURE 4. Urban exico:ini oefficientsor lternativencome
efinitions,
994-2006
Source:
uthor'slaborationasedn NIGH
various
ears).
Figures
and 5 show theevolution f theGini
coefficientsorurban nd
rural reas
n
Mexico for he 1994-2006
period.8
he
divergence
n
the
pat-
terns f
nequality y
sector s
quite striking.
n one
hand,
ncome
nequality
in urban reas n
Mexico,
regardless
f
the ncomedefinition
sed,
has
steadily
declined ince 1994. On the
other,
ncome
nequality
n rural reas
ncreased
until
000,
according
o thetotal ncome
definition,
r until
002,
according
to
any
other
ncomedefinition. fter
eaching
ts
peak,
ncome
nequality
n
rural reas
basically
returnedo ts 1994 level. The existence f sucha differ-
entiated
attern
f ncome
nequality
n rural nd
urban reas somehow
ug-
gests
hat ifferentactors ouldbe
affecting
he
dynamics
n
those wo ectors
of theMexican
economy.
his dea is
explored
n
moredetail ater.
The istributionf
Monetary
ncome
n
Mexico
The distributionf
monetary
ncome
n
Mexico is now
explored
n moredetail
by ooking
t the
growth
ncidence urves
GICs)
suggested y
Ravallion nd
8. Please ote hathissnot
rural/urbanncome
nequalityecomposition
xercise.his
analysis
efers
nly
o he ncome
nequality
ynamics
ithinuralnd rbanreas nd oes
not
iscusshe ontributionf ach ector
o otal
nequality
n
Mexico.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
10/35
Gerardo
squivel
163
FIGURE5. Rural exico:ini oefficientsor lternativencome
efinitions,
994-2006
Source:uthor's
laborationasedn NIGH
various
ears).
Chen
2003).
These curves howthe
percent hange
n
per
apita
ncome
long
the ntire ncomedistribution
etween wo
points
n
time.
Figure
shows the
GIC for he ntire 994-2006
period
t the
national, rban,
nd rural evels.
The
negative lope
in thefirst
raph learly
hows
why
Mexico's income
inequality
iminished
uring
his
period:
ncome t the
bottom
art
f thedis-
tribution
rew
faster han ncome
t themiddle nd the
op parts
f thedistri-
bution. hefigurelso shows hedifferentatternsollowed ytheurban nd
rural ncomedistributions
uring
his
period.
n
urban
reas,
ncome
growth
was
pretty
lat cross the entire istribution
xcept
for
he
top
three
eciles,
which
xperienced
maller nd in some
cases even
negative
ncome
growth
rates.
n
thecase of rural
reas,
two
aspects
re
salient:
irst,
verage
ncome
growth
was
greater
han
n
urban reas
(an
effect
hat,
iven
the
relatively
large
ural/urban
ap,
reduces
nequality),
nd
second,
he ural
GIC curve lso
has a
negative lope,
so that hebottom alfof the rural
ncomedistribution
had
higher
ncome
growth
ates han he
top segment
f
thedistribution.
ll
thesefactshave contributedo thereductionn income nequalitynMexico
that as taken
lace
since 1994.
Interestingly,
hese
results
lready uggest
hat hereductionn
nequality
in Mexico between
1994 and 2006 came fromdifferentources: in
urban
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
11/35
164
ECONOMIA,
all 011
FIGURE 6.
Mexico:rowthncidenceurves
sing onetary
ncome,
994-2006
Source:
uthor's
laboration
ased
n
NIGH
various
ears).
areas,
t was the result f the relative
and
for ome even
absolute)
oss of
income at the
top part
of the
distribution,
hereas
n rural
reas,
t
was
the
generalized mprovement
n rural ncomes as well as the
specific mprove-
ment n the ncome of the
relatively oor
ruralhouseholds
hroughout
his
period.
DecompositionAnalysis
f Sourcesof ncome
nequality
n Mexico
A
decomposition
f the Gini coefficients conductedbelow for the
years
1994, 2000,
and 2006 to
investigate
he contributionf differentncome
sources o theobserved
nequality
f
monetary
ncome n Mexico.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
12/35
Gerardo
squivel
165
Sourcesf
Monetary
ncome
The main
component
f
monetary
ncome n
Mexico is labor
ncome,
which
accounted or round 0
percent
f all
monetary
ncome
n
2006;
the econd-
largest
ource of
monetary
ncome
n
Mexico is income obtainedfrom he
businessesof
self-employed
ndividuals,
which ccountsfor
nother 0
per-
cent
of
monetary
ncome.
The
rest
f
monetary
ncome
proceeds
from vari-
ety
of
sources,
ncluding
ransfersnd remittances.
Table
A1
in the
ppendix
hows the
percentage
f households hat eceive
incomefrom ourcesother han abor ncome.This table shows thedramatic
increase hat as taken
lace
since
1992
in
the
percentage
f Mexican
house-
holds thatreceive some
type
of transfer.
lthough
ess
than
24
percent
f
households eceived
public
or a
privatemonetary
ransfer
n
that
year,by
2006
more
han 5
percent
f all
households
eported eceiving art
f their
monetary
ncome
hough private
r
public
transfer.he
single
most
mpor-
tantcontributoro this trend s
undoubtedly
he social
program rogresa
Oportunidades,
hich,
ccording
o 2006 ENIGH
data,
s received
by
15
per-
cent of Mexican households.9Two other
factors ccount for
part
of the
increase n transferso Mexican households:first,he ruralprogram ro-
campo,
whichwas intended o
support
ural
producers uring
he
transition
to trade iberalization
n
agricultural roducts,10
nd, second, remittances,
which re now received
n
7
percent
f Mexican
households,
wice as
many
as
in
1994. Based on
what s known
bout
the
distributiveffects
f the
Pro-
campo regressive)
nd
Progresa/Oportunidadesvery rogressive) rograms
(Esquivel
et al.
2010),
it is
quite ikely
that
hey
an
actually
ccountfor
great
deal of the
up-and-down ynamics
f ncome
nequality
n
rural reas
depicted
n
figure
.
Methodology
Lerman nd Yitzhaki
1985)
showed hat he
Gini
coefficientor
otal
ncome
inequality
G)
with
K
income ources an be
expressed
s
(1)
G
=
tstGkRk,
k=l
9. Formore etailsbout his
rogram,
ee Corbachond chwartz
2002)
nd
Levy
(2006).
10.
For
more etailsn
Procampo,
eeCorbachond chwartz
2002).
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
13/35
166 ECONOMI
,
Fall 011
where
k
s the hareof sourcek ntotal
ncome,
Gk
s theGini coefficientf
the ncome ource
k,
nd
Rk
s theGini
correlationetween he ource
ncome
k
and total ncome.
This
decomposition
f the Gini coefficient as a neat and
clear-cut nter-
pretation
ince t shows that
hecontribution
f
ncome sourcek to
inequal-
ity
depends
on
the nteractionf three lements:how
important
he ncome
sourceon total ncome s
(
Sk
;
how
unequally
distributedhe ncome source
is
(Gk);
and how
correlated he ncome
source and the distributionf total
income re
Rk.
Therefore,
n
income ourcethat
epresents
relative
arge
shareof total
income
ould have
a
large
effect n
inequality
s
long
as it s
unequally
dis-
tributed
that
s,
f t
has a
relatively igh
Gk
.
However,
f
Gk
s
low,
this
fac-
torwill dwarf he
ontributionf that ncome ource.
On theother
and,
f
an
income
ource s
veryunequally
distributedut s not
highly
orrelated ith
total ncome
as
in
the ase of
well-targeted
ransfer
rograms),
hen he on-
tributionf such a sourcecould n
factbecome
negative.
Stark,
aylor,
nd Yitzhaki
1986)
showed thatwith his
ype
of decom-
position
one
can estimate he
effect f a small
percentage hange n)
in
a
given
ncome source on
total
nequality holding
all other
ncome sources
constant)
hrough
he
following
xpression:
(2)
||
=
St(Gt*t-G)
or,
alternatively,
(3)
_Ak=^L_S^
This
expression
meansthat he
percent
hange
n
inequality esulting
rom
marginal ercentage
hange
n
ncome ourcek s
equal
to the nitial
hare f
income ourcek on total ncome
nequality
minus he nitial hare f
ncome
sourcek.
Gini
ecomposition
esults
Now
the
monetary
ncome
Gini coefficientsor
Mexico
are
decomposed
fol-
lowing
the
approach ust
described nd
using
the ncome sourcesdescribed
in
figure
and
tableAl. For
simplicity
f
exposition,
nstead f
applying
he
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
14/35
Gerardo
squivel
167
methodology
o thewhole
period
under
nalysis,
t s
applied
only
to the ur-
veys
of
1994,
2000,
and 2006.
The
descogini
Stata
command
presented
n
Lopez-Feldman
2006)
is used
in the
decomposition
xercise.
The
marginal
ffects
f the
decomposition
xercise
re shown
n
figure
.
Results
are
unequivocal:
at the national
evel,
there re three
nequality-
augmenting
nd three
nequality-reducing
ources
f ncome.
Among
hefirst
group
re
pensions,
ncome
from wn
businesses,
nd
ncome
from
roperty
rents.
Among
he econd
group
re ncome
abor
at
least since
2000),
remit-
tances, nd
transfers.
n
the
ast two
cases,
the
marginal egative
ffects
n
theGini
coefficient
ave
increased
hroughout
he
period.
Figure
7 also shows
the
marginal
ffect f
thedifferent
ncome ources
n
urban
nd rural
reas.
The
sign
of
the
marginal
ffects
f thedifferent
ncome
components
s
basically
he ame
as that bserved
t
thenational
evel.
There
are,
however,
ome
mportant
ifferences
n terms f the
relative
mportance
of
the
mpact
f some
sources.
For
example,
abor
ncome
s a
very
mportant
inequality-reducing
orce
n urban
reas but
not
n
the
rural ector
there,
t
even
augmented
nequality
n
2006).
On theother
hand,
ransfers
re a
very
importantnequality-reducingactor n rural reas butnot as importantn
urban
nes.
Finally,
note that
emittances
o not seem
to
have a
large nega-
tive
marginal
ffect
n
nequality
n
any pecific
ector,
lthough
hey
re rel-
evant
t thenational
evel.
This
apparent
aradox
s
explained
by
the
fact
hat
whiletheGini
correlation
f remittances
ith ural
monetary,
ncome
s close
to 50
percent,
hey
ave
a much
owerGini
correlation
ith
monetary
ncome
at the
national evel.
In that
ense,
remittances
ave
an effect
t the
national
level because
they
re
heavily
oncentrated
n
thebottom
alf
f the
national
income
distribution.11
herefore,
emittances
ork
s an
inequality-reducing
sourceof incomethroughherural/urbanncomegap and notthroughhe
sector-specific
ncome
distribution.
Why
abor
ncome
Has Become
an
Equalizing
ncome
Force
The results
f the
decomposition
xercise
uggest
hat ne
of the
most
mpor-
tant
qualizing
orces
n
recent
ears
n Mexico
has been
the volution
f
abor
income,
oth
n urban
reas and
n the
country
s a
whole.
n
fact,
hereduc-
tion n thetotal ontributionf abor ncometo theGinicoefficientccounts
for
almost all
of the observed
reduction
n this coefficient
hroughout
he
11.
For
more etails
n hisssue
ee,
or
xample,
squivel
nd
Huerta-Pineda
2007).
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
15/35
FIGURE 7.
Marginal
ffectnGini oefficient
y
ncomeource:
verall,rban,
and
Rural
exico
Source:
uthor'slaboration
ased
n
NIGH
various
ears).
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
16/35
Gerardo
squivel
169
1994-2006
period.
Therefore,
nderstanding
he nature f the
change
n
the
effect
f abor ncome
on
inequality,
hichwentfrom
eing
positive
n
1994
to
becomingnegative
n
2000
and 2006
(see
figure
),
is crucial to under-
standing
hewhole
dynamics
f ncome
nequality
n
Mexico
since
1994.
To
begin
with,
ote hat abor
ncome
s
basically
the
product
f
multiply-
ing
hourly
wages by
number
f hours
worked.That
being
the
case,
leaving
aside
changes
n the number
f hoursworked
long
the ncome distribution
(which
ould
have occurred
ut
probably
ot
necessarily
n
the
magnitude
r
directionhat ouldactually xplain heobserved hanges n ncome nequal-
ity),
he
only
other hannel
hrough
which abor
ncome can affect
ncome
inequality
s
through
hanges
n
wage
rates.
Therefore,
ost f the
hanges
n
this
type
of
inequality
omehow
mustbe
the outcome
of
changes
n
wage
inequality.
n some
sense,
this s a
very
fortunate
ircumstance,
ince
a link
can then
e established
etween
hisdiscussion
n income
nequality
nd
the
literature
n
wage inequality
n Mexico that
has been
written s
part
of the
debateon
the
relationship
etween
rade
nd
wages.12
Let us
look first
t the evolution
f
wage inequality
n
Mexico
in
recent
yearsby usinga standard efinitionfwage inequality ivenbythe ratioof
the
wages
of
nonproduction
orkers o
those f
production
orkers.
his ratio
is also
(grossly)
defined
s the
skilled/unskilled
age
ratio,
wherein
onpro-
duction
workers re
considered
proxy
or killed
abor nd
production
ork-
ers a
proxy
or
unskilled
abor.13
Figure
8 shows
theevolution
f
thismeasure
f
wage
inequality
n Mexi-
can
industry
or he
period
from 984
through
007.
The data
for his
graph
come
from he
Encuesta
ndustrial
Mensual
(EIM)
[Monthly
ndustrial
ur-
vey],
which
has
monthly
nd annual
data
on total
wages
paid
and
totalhours
workedn ndustry ybothproductionndnonproductionorkers. hisfig-
ure is an
updated
version
of similar
versions
published
n,
for
example,
Esquivel
and
Rodriguez-Lopez
2003)
and
Chiquiar
2008).
The
pattern
f
wage inequality
n Mexico's industrial
ector
n
figure
is
remarkably
imilar o
the volution
f
nequality
nder
hevarious
definitions
of income
thatwere
shown
before.
This
figure
hows a continuous
pward
increase
n
wage inequality
ince 1984
that
asted until
he
mid-1990s,
fol-
lowed
by
a
steady
decline since
then.
A
slight
ifference
etween
his
graph
12.
See
the bundant
eferences
o heMexican
ase hat
ppear
n
Goldberg
nd
avcnik
(2007),
survey
n
globalization
nd
nequality.
13.
This
s,
of
ourse,
gross
implification,
ince here
re
roduction
orkers
ho re
highly
killednd
onproduction
orkers
ho
re
elatively
nskilled.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
17/35
-
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
18/35
-
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
19/35
172
ECONOMI
,
Fall
011
forty-six
fthe
forty-eightanufacturing
ranches. etween 1994 and
2007,
however,
he
pattern
f the skilled/unskilled
age
ratio
n
Mexico's manu-
facturingndustry
ooks somewhat ifferentnd more
heterogeneous
han
n
the
previous eriod:
most ndustries
ow show
a
slightly ecliningwage
ratio
between hese wo
years,
ut here lso are a fewbranches
n which
he
wage
ratio
s
now either he ame or
slightly
bove
its
1994
level.
On theother
and,
data on the
volution f the killed/unskilled
age
ratio
at the state
evel
show
also a
clearlydeclining
rend
n
almost
very
tate
n
Mexico
since
hemid-1990s
Esquivel2008).
In
summary,
ince
1996 there
as
been an
important
eduction
n
wage inequality
n
Mexico. This reduction as
taken
lace
not
only
t the
ndustry-wide
evel but lso in most
manufacturing
branches nd across the
country
n
manyregions
nd states.
Consequently,
a
good explanation
f labor ncome
nequality
and
of
wage inequality)
as
to be able to
explain
not
only
the
rapid
ncrease
n
wage inequality
etween
1984 and
1996
but lso the
eduction
n
wage
nequality
hat as beenobserved
since
1996.
Explaining
he
Evolution f
Wage Inequality
n Mexico
The
rapid
ncrease
n
wage inequality
hat ccurred n
Mexico between1984
and
1994
or
1996 has been
widely
documented nd studied.15 n
interesting
aspect
of this
rend
s
that ts
beginning
oincidedwith heunilateral
pening
of theMexican
economy
hat
tarted
recisely
n the
mid-1980s. he increase
in
Mexico's
wage
inequality
herefore ould be somewhat
nexpected,
on-
sidering
hatMexico has a relative
bundance f unskilled abor
at
eastfrom
theperspectivef tsmaintrade artner,heUnited tates)and that tandard
trade
heorieswould have
predicted xactly
he
opposite
pattern
that
s,
a
reduction
n
the
killed/unskilled
age
ratio;
ee
Cragg
nd
Epelbaum
1996).
As
a
consequence,
everal
ossible
hannels
most
fthem inked o the
pen-
ing
of the
economy
n
the
mid-1980s)
have been
suggested
o
explain
this
apparent aradox.
The
explanations
hathave been
proposed
to
explain
the
post-openness
increase
n
Mexico's income
nequality
an be
grossly
ivided nto wo
groups:
in
the
first,
he
xplanations
mphasize
actors
ffecting
hebottom
art
f the
15.
See,
for
xample,squivel
nd
Rodriguez
6pez
2003),
Airola ndJuhn
2005),
Robertson
2007),
costand
Montes-Rojas
2008),
hiquiar
2008),
erhoogen
2008),
nd
the eferences
ited
herein.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
20/35
Gerardo
squivel
173
incomedistribution
that
s,
the
egment
hat or hemost
part
omprises
ess
skilled
nd
ess
experienced
orkers);
he econd
emphasizes
actors
ffecting
the
upperpart
f thedistribution.
n
thefirst
roup,
or
xample,
re theories
emphasizing
hereduction
n
real minimum
ages
Fairris,
opli,
and
Zepeda
2008)
as well as theories
uggesting
hat he
mid-1980s
reduction
n
tariffs
disproportionately
ffectedndustrieshat
mployedmostly
ow-skilled
work-
ers
(Hanson
and Harrison
1999).
In
the second
group,
ome
explanations
emphasize
heroleof an increase
n thedemand or
killedworkers
ssociated
withthepresenceof foreignnvestmentFeenstra nd Hanson 1997);
with
skill-biased
echnological hange Cragg
and
Epelbaum
1996;
Esquivel
and
Rodriguez-Lopez
2003);
and
with a
quality-upgradingrocess
due to an
increase n
exports
Verhoogen
2008).
Other
explanations
have
suggested
that
ducation
nequality
ould have also
played
role
Lopez-
Acevedo
2006)
or that these trends ould
indicate
only
short-run ffects
Canonero
and
Werner
002).
On
the
other
hand,
he
post-
996 reduction
n
wage inequality
n
Mexico
has been much less studied. So
far,
only
Robertson
2007)
and
Campos
(2008) haveanalyzed his rend.WhileCamposfavors n explanation ased
on
supply
actors,
obertson
uggests
hatMexico's
manufacturing
orkers
are now
complements
f
rather han substitutes or U.S.
workers nd that
there as been an
important
xpansion
f
assembly
ctivities
n
Mexico
that
has increased
hedemandfor ess
skilledworkers.
Of
course,
many
f the
proposed
xplanations
or he
pre-NAFTA
ncrease
in
wage inequality
n
Mexico
are not
mutually
xclusive,
nd
they
ould
in
factbe at least
partially
orrect.
However,
t is also true
hatmost of them
cannot
explain
the
subsequent
eduction
n
wage inequality
hathas been
observed since 1996. Therefore heseexplanations re either ncorrect r
incomplete,
ince there
ould be
manyunderlying
orces
cting
n
different
directions.
hat is
why
Robertson
2007)
noted thatthe
pattern
f
wage
inequality
n Mexico is
puzzling
because no
singletheory
ould
explain
the
evolution
f
wage inequality
efore
nd afterNAFTA.16
Although
t is not
the
objective
of this
paper
to
identify
r to establish
which
xplanation
if
any)
s
correct,
t east some of them ould
be ruled ut
16. There
re,
owever,
ome entativeheoretical
xplanations
or uch
pattern.
or
example,tolia2007) uggestedhat,nderertainircumstances,venf he tandardre-
dictionromHeckscher-Ohlin-Samuelson
odel orkss
predicted
n he
ong
un,
here
ay
be ome hort-run
or
ransitory)
ffectsf radeiberalization
hatead o differentutcome
because f wo actors:
irst,
n
symmetry
n he ontractionnd
xpansion
f ome ectors
and, econd,
he
apital-skillomplementarity
n
production.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
21/35
174 ECONOMI
,
Fall 011
by
ooking
t some
wage
data
provided y Campos
2008).
The next wo
fig-
ures show
the mean
og wage
of male workers
n
Mexico
for elected
years
and fordifferent
ombinations f education
nd
years
of
experience.
Work-
ers reclassified
ccording
o the
evel ofeducation chieved
less
that ower-
secondary,
ower-secondary,pper-secondary,
nd
college
education)
nd to
the number f
years
of work
experience
less
or more
than
wenty ears
of
experience).
The
upperpart
f
figure
0
showsdata for he
years
1989, 1994,
and 1996
and the ower
part
shows information
or
1996
and 2006.
The first
igure
shows
an
interesting
esult: etween1989
and
1994,
mostof the
changes
n
the
wage
distribution
n Mexico occurred
n
the
upper
ail of thedistribution.
That
s,
the ncrease
n
wage inequality
n
those
years
annot
e
explained
by
a reduction
n
the
wages
of ow-skilled r
nexperienced
orkers;nstead,
he
increase an be
explainedonlyby
an increase
n
the
wages
of
highly
killed
or
highly xperienced
workers. his result
asically
rules ut
ny xplanation
based on
changes
n
the ower tail of the
wage
distribution,
uch as those
based on a
falling
eal minimum
wage
or on a biased
openness
of unskilled
labor-intensive
ndustries. his
figure
lso shows the
widespreadnegative
effects f thefinancial risis
of
1994-95,
which
reduced,
lmost
proportion-
ally,
he eal
wages
of
all
types
fworkers
n
Mexico between 994
and
1996.
The bottom
art
of
figure
0 shows the
wage
distribution
n
Mexico for
1996
and 2006. Unlike
figure
,
thisone shows thatmostof the
changes
n
the
wage
distributionook
place
in
the ower ail.That
s,
workerswith ower
levels of education nd/or ewer
years
of work
experience
had the
argest
increases n their
verage wages,
and that
xplains
the reduction
n
wage
inequality
hat as been observed ince
1996.
This also
suggests
hat
ny
on-
vincing tory f thepost-NAFTAreductionnwage inequality as toexplain
the ncrease
n
the
wages
of
ow-skilled/less
xperienced
workers ather han
thereduction
n the
wages
of
high-skilled/morexperienced
workers.
The
previous
esults
onfirmhe ntuition hat here s no
singleexplana-
tion
for he
evolution
f
wage inequality
n
Mexico since
1984.
Indeed,
the
fact hat
he
1984-94
increase
n
wage inequality
s associatedwith
hanges
in the
upper
ail of thedistribution hilethe
post-NAFTA
reductionn
wage
inequality
s associated
mostly
with
hanges
n
thebottom
ail
suggests
hat
there re at east two
eading
forces t
play.
n
thefirst
ase,
as
discussed,
he
only xplanationshat eemtobecompatiblewith he bserved rend re those
suggesting
he
presence
of skill-biased
echnological hange,
ither
xoge-
nous
change Cragg
and
Epelbaum
1996
and
Esquivel
and
Rodriguez-Lopez
2003)
or
endogenous hange resulting
rom he
presence
of multinational
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
22/35
Gerardo
squivel
175
FIGURE 10. Mean
ogWage
fMale
Workers
y
ducationnd
xperience,
elect ears
Source:uthor'slaboration
asedn
ampos
2008).
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
23/35
176 ECONOMI
,
Fall 011
firms
Feenstra
nd Hanson
1997)
inMexico and/or
y
the
upgrading
f the
quality
f
exporting
irms
Verhoogen
008).
For the
post-NAFTA
period
here re at
east
three
ossible explanations,
two of whichhave
already
been mentioned nd are
not
mutually
xclusive:
an increase
n
the
supply
of
relatively
killed
workers
Campos
2008)
and
an
increase
n the demandforunskilled abor
resulting
rom n
expansion
of
assembly
ctivities
n
Mexico's
manufacturing
ector
Robertson 007).
Either f thetwo
effects ould
explain
the
reduction
n the
killed-wage
re-
mium bserved
n
the data.
A
third
xplanation
hat s also
compatible
with
the
previous
wo
s
that f
a
standard eckscher-Ohlinffect
n a
country
uch
as Mexico
in
whichunskilled abor s abundant
Chiquiar
2008).
This
effect
could
be the
ate outcome
f
trade iberalization
uggested y
Canonero nd
Werner
2002)
and
already
modeled
by
Atolia
(2007)
or,
alternatively,
n
underlying
ffect hat id
notshow
up
in
the
data
before ue to the
presence
of a
stronger
orce,
uch as a skill-biased
echnological
hange
s
previously
hypothesized y
Esquivel
and
Rodrfguez-L6pez
2003).
A
much more
detailed and
rigorous nalysis
s needed to discriminate
among
hese lternative
ypotheses. owever,
t s
possible
to
move
forward
by
analyzing
whetherome
of
these
hypotheses
re
borne
ut
by
thedata.
Fig-
ure
11
shows he
omposition
fMexico's workforceetween
989
and 2006
according
o the evels of education nd
experience
efined bove. This com-
position
bviously
eflectshe
nteractionf
both
upply
nd demand actors.
In
general,
he
figure
hows
that
hroughout
he
period
therewas a
large
reduction
n
the hare fthe east skilled nd
east
experienced
workers
those
with ess than
econdary
ducation)
nd an increase
n
the hares
f
theother
types
of workers.
he most dramatic
hanges,
however,
ook
place
in the
shareof workerswith ess thansecondary ducation. n fact,thisgroup,
which ccounted or lmost
5
percent
f theworkforce
n
1989,
represented
only
bout one-third f the workforce
y
2006,
a reduction f about
20
per-
centage oints
n
a
seventeen-yearpan.
Thatreduction as
compensated
or
by
increases
n
the shares
of
most
of
the other
groups
of workers. hese
trends,
whichwere
alreadypresent
etween
1989
and
1994,
accelerated
n
the
post-NAFTAperiod.
These results herefore
uggest
hat t east
part
f
the
relative ncrease
n
the
wages
of the ow-skilled/less
xperienced
workers
n
Mexico is associ-
ated with hechange nthecomposition ftheworkforcend, nparticular,
with
reduction
n
the number f unskilledworkers ather han n increase
in
the
upply
f skilledworkers. f
course,
his
result
s not at
all
incompat-
ible
with he
hypothesis
hat he
demand
for
unskilledworkers
ncreased,
s
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
24/35
Gerardo
squivel
177
FIGURE 11. Workforce
omposition
y
ducationnd
xperience,
989-2006
Source:uthor'slaborationased
n
ampos
2008).
suggested
n Robertson
2007),
but,
by
itself,
he result annot
explain
the
simultaneousncrease n therelative
wages
and reductionn the
participation
of these
workers
n Mexico's totalworkforce.
Finally,figure
2 shows some results hat re
compatible
withthe
view
that mphasizesthe role of thecomposition f the abor force.The graph
shows on the x-axis the
change
between 1996 and 2006
in the share of the
eight
different
roups
of workers
ccording
to their evels of education
and
experience
s defined bove.
Participation
n Mexico's workforce as
declined n three
groups,
which
correspond
o the east educated and less
experienced
workers. he
y-axis
ndicates he
verage hange
n the
og wage
of male and femaleworkers hat
elong
to each of the
groups.
As
expected,
the
groups
whose shareshave diminished n the
past
decade are those that
have had the
argest
ncrease n
wages.
Notice
that he ncreases
n the
wages
of these workers re close to 20 percent and in some cases even close to
30
percent throughout
he
en-year eriod.
On the ther
and,
he
ategories
of workerswhose shares n Mexico's workforce ave increased
the
more
educated/more
xperienced
workers)
ave tended o have either
tagnant
r
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
25/35
178
ECONOMIA,
all 011
FIGURE 12.
Change
n haref otalWorkers
y
ducationnd
xperience
ersus
hange
in
ogWage y
Gender,
996-2006
Source:uthor'slaborationasedn
ampos
2008).
even
decreasingwages
since
1996.
This
graph
hen
upports
he
hypothesis
that he
hange
n the
omposition
fMexico's workforces the
eading
force
in the
reduction
n
wage
and labor ncome
nequality
n
Mexico
in
the
post-
NAFTA period.
Summary
nd Conclusions
This
paper
reviews he
pattern
f income
nequality
n Mexico since
1994,
when NAFTA went nto
effect.
Using
information
rom
nationally epre-
sentativehousehold
surveys,
t shows that there has been an
important
reduction n income
inequality
ince
1994
and thatthis trendhas almost
reversed ncome nequality o the evels thatwere observedbefore herapid
increase n
inequality
hat ook
place
between
1984
and 1994.
As shown
by
a Gini
decomposition nalysis by
income
source,
abor
income, remittances,
nd
public
transfers
mainly through
he
Progresa/
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
26/35
Gerardo
squivel
179
Oportunidades rogram)
ave all
played
an
important
ole n this
qualizing
process.
n
particular,
he
paper
shows that abor ncome
has become a
very
importantqualizing
force
n
urban reas
in
Mexico,
while
public
transfers
have been
especially
important
n
reducing nequality
n
the
rural sector.
Remittances,
n the
other
hand,
have been a national
nequality-reducing
source of ncome since 1994.
The
paper
also
provides
ome evidence
suggesting
hat heforces hat ed
to a
sharp
ncrease
n
wage inequality
cross all industries
n
Mexico
during
the
1980s
and
early
1990s
are no
longer operating.
n
fact, generalized
reduction
n
wage
inequality
cross
ndustries nd
regions
n
Mexico
is now
observed,
uggesting
he
growing
elevance
f other lements
n
this rend.
In
general,
believe that Mexico
is now
beginning
o
experience
the
inequality-reducing
ffects f
having
more ducatedworkforcend
of trad-
ing
with
more
kill-abundant
ountries.
his
equalizing
effect eems
to have
been
postponed y
a skill-biased
echnological
hange
either
xogenous
or
endogenous)
r
by
an
endogenous
echnological pgrading
hat,
n
any
case,
now seems
to have ended.
This fact nd an ambitious
nd
widespread
ocial
program ocusedon poorruralhouseholds eemto be the mainexplanatory
factors
n
the
harp
eduction
n
inequality
hathas been observed
n Mexico
in recent
ears.
Appendix
TABLE A Households
hat eceivencome
romourcestherhan abor
ncome
Percentf otalouseholds
Sourcefncome 1992 1994 1996 1998 2000 2002 2004 2005 2006
Ownusiness
43.8 42.7 43.3 43.2
41.1 41.9 38.1
39.1 42.0
Property
ent
4.4 3.5 3.7 3.5
3.1 4.0 4.7
4.0 4.6
Financialncome
25.9 15.2
22.9 19.1 19.1
19.3 20.2 18.1
23.5
Transfer
23.5 23.8
29.0 31.2 34.0 38.6 42.0
41.3 45.5
Remittance
3.7
3.4 5.3 5.3 5.3
5.7 5.6 6.0
7.0
Pension
8.8
8.1 8.2 9.5 10.0
10.0 11.8
11.4 11.9
Publicnd
rivate
ransfers
3.6 14.5
18.7 19.6 23.1 28.4 31.2
30.5 34.6
Procampo
... 1.2
4.6 2.8 2.6 5.6
4.6 3.5
4.0
Progresa/Oportunidades
.a.
n.a. n.a. n.a. n.a.
12.3 13.4 13.5
14.8
Source:uthor'sstimatesasedn
NIGH
various
ears).
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
27/35
Comment
Guillermo
Cruces:
The
paper y
Gerardo
squivel
n this dition fEconomia
presents
thorough
ccountof
inequality
rends
n Mexico since
the mid-
1990s.
Papers
uch s
this
onstitute
fundamentalxtension
f thediscussion
of
aggregate egional
rends
uch as that
rovided
y Gasparini,
ruces,
nd
Tornarolli
201 1),
and
they resent
n
opportunity
o
discuss
n
depth
he
fac-
tors
nderlying
he
volution
f
nequality
t the
ountry
evel.
dentifying
he
fundamental
auses of the volution
f a national
ncomedistribution
s never
aneasyfeat, nd the uthormust e laudedfor roviding coherentnd com-
pact
discussion f
an eventful
eriod
n
Mexico
that ncluded
major
ransfor-
mations
with
potentiallyarge
effects
n the ncome
distribution,
ncluding
macroeconomic
rises,
market-oriented
tructural
eforms,
he
opening
f the
economy
o nternational
apital
flows,
free
rade
greement
ith
heworld's
largest conomy, echnological
hange,
ncreased
globalization, political
transition,
nd even
an armed
ndigenous
prising.
This comment xamines ome
evidenceon distributional
hanges
n
terms
of
poverty
eduction o
complement
he article's
main focus on
inequality
and then iscussesthe uthor's onclusionsnterms f a structuralhange n
inequality
rends
n
Mexico
in
light
of the effects f the
nternationalco-
nomic risis hat tarted
n
2007.
Finally,
t
suggests
ome avenues
for urther
research.
The first
oint
f
this omment oncerns he
evolution f
poverty
verthe
period
f time overed
by
the rticle.National
ggregate
overty
ates
or he
US$2.50
and
US$4.00
purchasing
ower parity
PPP)
international
overty
lines re
presented
n
figure
for he
period
1989-2008
and also
for heurban
and rural
opulations
or he
US$2.50
poverty
ine.This
figure rovides
nfor-
mation n a dimensionfdistributionalhange hat omplementshe horough
discussion
f
the evolution
f
inequality
n the
paper.
The
figure
ndicates
Guillermoruces
s with
EDLAS-UNLP
nd ONICET.
180
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
28/35
Gerardo
squivel
181
FIGURE 1.
Poverty
ate or ationalndRural/Urban
opulations,
exico,
989-2008
Source:EDLASnd orldank
2011).
asednMexico'sNEGIiannual
urvey,
ncuestaacionale
ngresos
Gastose os
Hogares
ENIGH).
moderate all
in
poverty
n
the
early
1990s,
followed
by
a
large
increase
between1994 and 1996
that an be
attributed
ainly
o the macroeconomic
crisis f 1995.
From
1996
until
006,
there s
a
continuous all
n the
poverty
ratespresentedn thefigure.n fact, hisevidence ndicates hat he notable
results rom
igure
in
the
paper,
which how substantial
rowth
n
ncome t
the
bottom f thedistribution
n the
period
1994-2006,
could be even
arger
f
computed sing
he
post-crisis ear
1996
as a basis.
This discussion llustrates furtherimension f the
distributional
ynam-
ics documentedn
figure
of the
paper:
thereductionn
nequality
n Mexico
was
accompanied y
a substantial all n the
proportion
f the
population
iv-
ing
underdifferent
overty
ines and reflects he ncrease n standards f
living
nd the
pro-poor
ature f the
growth rocess
n theMexican
economy
over theperiodunder tudy.At thesame time, hisevidence also indicates
that here s still
ample
room for
mprovement:
bout 15 and 30
percent
f
the
population
still lived under the
US$2.50
and
US$4.00
poverty
ines,
respectively,
n
2008,
and
major egional nequalities
re evidentn the
poverty
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
29/35
182
ECONOMIA,
all 011
FIGURE 2. Laborncome
overty
rendndex
nd
early
DP rowthate1
Source:
0NEVAL(2011).1heLabor
ncome
overty
rendndexndicateshe
roportion
fndividuals
ho
annotoverheostfbask
foodasketithheiraborncome.ts asednMexico'sNEGI
uarterlyurvey,
hencuestaationale
cupaddnmpleo
ENOE).
a. Fourth
uarter
ith
espect
o ourth
uarter
fhe
reviousear;
aseourth
uarter
f 0051
levels n rural
reas,
which re about wicethenational
ggregate.
While the
paper
discusses urban-rural
ifferentials,
roviding
more evidence on their
nature nd discussing heir onsequences n terms f policy are certainly
worthwhile irections orfurtheresearch.
inally,figure
reveals a small
but
significant
ncrease n
poverty
etween 006 and
2008;
the
mplications
of this
hange
nd its
underlying
actors re discussednext.
The second
point
f this omment efers o the
ustainability
f thedistri-
butional
hangeprocess
documentedn the
paper
nd to ts
vulnerability
ith
respect
o
aggregate
hocks.The 2010 nationalhousehold
urvey
ENIGH)
was not vailable at the ime his ssue went o
press, mplying
hatno
aggre-
gate poverty
ates
comparable
o those n
figure
were available.
Figure
2
presentsn alternative easure, heLabor ncomePoverty rend ndex con--
structed
y
the
Consejo
Nacional
de
Evaluaci6n de
la Polftica e Desarrollo
Social
[National
Council for Evaluation of Social
Development Policy]
(CONEVAL)
and
based on
INEGI'
s
quarterly
urvey,
heEncuestaNacional
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
30/35
Gerardo
squivel
183
de
Ocupacion
y Empleo
[National
Survey
of
Occupation
nd
Employment]
(ENOE).
The
index,
which ndicates
he
proportion
f ndividualswho can-
not cover
the
cost
of a basic
food basket with heir abor
ncome,
has been
normalized o
1
forthe fourth
uarter
f
2005.
The evidence
n
the
figure
illustrates
series
of
factors.
irst,
he izable
growth
n
GDP
in
2005, 2006,
and 2007
(between
and
4
percent
ach
year)
did nottranslate
nto ubstan-
tial reductions
n
aggregatepoverty.
n
fact,
therewas a
relatively
arge
increase
n
urban
poverty
atesfor
2005-07,
which can be attributed
o the
increase nthepriceof foodrelated o anupward rendncommodity rices
and
growth
n the
global
economyduring
hat
eriod.
Second,
the
figure
ot
only provides
furthervidence on the ncrease
n
aggregate overty
etween
006 and 2008 but also
illustrates he
very arge
distributional
mpact
of the nternational
conomic
crisis,
withthe
poverty
rate
ncreasing
urther
n
the2008-10
period.
That ncrease
was seen
even for
therural
population,
which
was
relatively pared
from he 2005-07
upward
trend.
The
large
ncrease
n
poverty
or 2008-10 is all
the more
worrying
when
considering
hat he
nternationalrisis had
only
a
relatively
modest
effectat leastbyLatinAmerican tandards) n domesticgrowth ates,with
reductions
f about
1
and
2
percent
f
GDP
for 008 and
2009,
respectively.
Moreover,
he
reversal
n
this
trend etween2009
and
2010,
with
yearly
growth
ateof about
4.5
percent,
id nottranslate
nto further
eduction
n
poverty:
he
Labor Income
Poverty
Trend ndex increased
although
only
slightly)
ver
that
eriod.
The evidence
in
figure
1
points
toward a more
fundamental
uestion
related
o thenature
nd the
imits f the
process
of distributional
hange
doc-
umented
y
the
Esquivel
paper
for he
period
1994-2006.
While
inequality
andpoverty ellsubstantiallyvertheperiod, he reversal n growth rends
resulted
n a
large
increase
n
aggregate
poverty
evels
and,
apparently,
change
in the mechanism
inkinggrowth
with
poverty
eduction,
s wit-
nessed
by
the
tagnant
overty
evels of 2010.
Moreover,
heSEDLAC
data-
base
(CEDLAS
and World
Bank
2011),
on which
figure
in
the
paper
is
based,
indicates n
increase
of 0.7
in
the Gini
coefficient
etween2006
and
2008
(although
he
change
s not
statisticallyignificant
t standard
evels).
While
further
ains
n distributional
erms re
always
harder
o
accomplish
t
lower
poverty
nd
nequality
evels,
the
mpact
f the
nternational
conomic
crisis ouldprovide frameworkoadjust hepolicies mplementedver the
previous
period, specially
n terms f the
breadth,
epth,
nd
modality
f
safety
ets nd also
in
terms
f the
mplications
f the
productive
tructure
f
the
economy
nd ts
inks
with
major
economies.
This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/9/2019 Paper Income Inequality Dynamics Since NAFTA.pdf
31/35
184 ECONOMI
,
Fall 011
TABLE
1
Changes
n he killed
age
remiumndRelative
upply
f killedndUnskilled
Workers
ndRelative
emand,
exico,
989-2008
Relative Relative Relative
Wage
Relative demand demand
demand
Period
premium supply (a=2) (c=3)
(o=4)
1989-2000 1.8 3.6
7.2 9 10.8
2000-08
-2.