Sie sind auf Seite 1von 35

The Dynamics of Income Inequality in Mexico since NAFTA [with Comment]

Author(s): GERARDO ESQUIVEL and Guillermo Cruces


Source: Economa, Vol. 12, No. 1 (Fall 2011), pp. 155-188
Published by: Brookings Institution Press
Stable 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 support@jstor.org.

Brookings Institution Press is collaborating with JSTOR to digitize, preserve and extend access to Economa.

http://www.jstor.org

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

GERARDO
The

Dynamics

of Income
in Mexico

ESQUIVEL
Inequality
since

NAFTA

Mexico is a highlyunequalcountryis a factthathas been recognized


at least since Alexandervon Humboldtwroteat the beginningof the
Thatnineteenth
centurythatthe regionthenknownas New Spain was "the
countryof inequality."Sadly, thisis stilltruein thetwenty-first
century.For
Corbacho
and
Schwartz
out
that
"Mexico's
income
(2002) point
example,
is
more
inequality significantly
pronouncedthantheLatinAmericanaverage,
whichis theregionwiththehighestdegreeof inequalityin theworld."Also,
looking at the long-runtrendin income inequalityin Mexico leaves little
roomto be optimistic.Historicaldata show thatwhile Mexico achieved an
reductionin inequalityduringthe 1960s and 1970s,periodsof very
important
rapideconomic growth,the countryhas experiencedverylittleprogressin
incomedistribution
since the 1980s (Sz6kely 2005).
This situation,however,mayhave startedto changein recentyears.This
paperprovidesevidenceon thereductionin incomeinequalitythathas taken
place in Mexico since 1994 and discusses some of thelikelysourcesof this
trend(see figure1), whichis important
forat leasttworeasons:first,
because
ithas almostcompletelyreversedthewidelydocumentedincreasein inequalitythatoccurredin the1984-94 period(Bouill6n,Legovini,and Lustig2003;
Legovini,Bouill6n,andLustig2005), andsecond,becausethereductionseems
to be theresultoftwoimportant
structural
changesin theMexican economy:
Gerardo
iswith
El ColegiodeMexico.
Esquivel
Thispaperwasoriginally
written
as partoftheUnited
Nations
Development
Program
theState,
andtheDynamics
ofInequality,"
coordinated
"Markets,
(UNDP)project
byNora
andLuisFelipeL6pez-Calva.
I thank
inprevious
seminars
atUNoffices
in
Lustig
participants
NewYorkandMexicoCity,
as wellasparticipants
attheLatinAmerican
andCaribbean
Economic
Association
inRiodeJaneiro
andBuenosAiresfortheir
useful
(LACEA)congresses
comments
andsuggestions.
I amparticularly
thankful
toGuillermo
Cruces
andAdriana
Kugler
fortheir
comments
andrecommendations
ona previous
version
ofthiswork.
Edith
Cortes
providedtruly
research
assistance.
Theusualdisclaimers
outstanding
apply.
155

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

156

ECONOMIA,Fall2011

Gin!
Coefficient
andTheil
Index
FIGURE 1. Mexico's

Source:
CEMAS
and
Worfd
Bank
(2010).
thearrivalofbetter-targeted
social programssuchas Progresa/Oportunidades
and a reductionin laborincomeand wage inequalitythatseemsto be associatedwiththeimprovement
in educationallevels in Mexico. A thirdcontributin inequalityhas been thegrowingflowof
factor
to
the
recent
reduction
ing
remittances
thatmanyMexicanslivingabroadsendtotheirfamiliesleftbehind
in Mexico.
The possibilitythatbothsocial policyand educationalimprovements
partiallyexplainthereductionin incomeinequalityin Mexico cannotbe underestimated.In fact,incomeinequalityis diminishing
in severalLatinAmerican
countries,and it is possible thatsimilarfactorscould be at play in manyof
thosecountries.1
This could lead notonlyto an appropriateevaluationof the
new social policies thatare being implementedin the regionbut also to a
reconsideration
of theeffectthathigherlevels of education,combinedwith
1. See,forexample,
Ferreira,
Leite,andWai-Poi
(2007)andBanosandothers
(2010)for
andCruces
andEberhaid
andEngel(2008)
theBrazilian
case;Gasparini
(2010)forArgentina;
viewonrecent
trends
ininequality
inLatin
seetheintroforChile.Fora more
America,
general
duction
toLdpez-Calva
andLustig
andTornarolli
Cruces,
(2010)as wellasGasparini,
(2009).

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 157
TABLE 1. Overview
ofMexican
Economy
Feature

I
Stage
1950-70

II
Stage
1970-82

III
Stage
1982-94

IV
Stage
1994-2000

V
Stage
2000-06

crisis
and Low
growth High
growth Low
growth 1994
High
growth
macro- with
with
macro- with
macro- recovery;
low
with
macroeconomic
economic
economic growth
with
economic
inflation stability
stability
instability adjustment some
free
trade Open
economy
Openness Semidosed Semidosed Unilateral NAFTA;
with openness agreement
with economy
economy
tariff
and
in1985;
tariff
and
with
the
EU
nontariff nontariff beginning
barriers
barriers
ofNAFTA
negotiations
in1992
reductionIncrease
stable Rapid
Reduction Reduction
Inequality Relatively
Low
Low
of
Social
Nontargeted
Targeted
Expansion
social
in
programs
Progresato
protection
areas:
urban
programs: rural
Solidaridad Progresa
and
areas
Procampo
Author's
Source:
compilation.
Macro

globalizationandtradeliberalization,
mayhaveon inequalityin middle-income
countries.2
The paper firstprovides an overview of macroeconomicconditionsin
Mexico duringthepast decades followedby estimatesof incomeinequality
in Mexico using alternativedefinitions
of income.Next,a Gini decompositionanalysisis conductedto investigatethecontribution
of different
income
sourcesto theevolutionof inequalityin Mexico, and theroleof incomelabor
and wage inequalityis discussed in explainingthe dynamicsof inequality.
The conclusionsfollow.

AnOverviewof Mexico'sEconomicConditionssince 1950


Table 1 providesan oversimplified
summaryof Mexico's economicperformancesince 1950. In thefirststage(1950-70), GDP grewat a relativelyrapid
low fiscaldeficits,and
pace (3 percenta yearpercapita),withpricestability,
a fixed exchange rate since 1956. The second stage (1970-82) was again
2. SeeGoldberg
andPavcnik
onthisissue.
(2007)fora recent
survey

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

158

ECONOMI
A, Fall2011

a period of rapid growth(3 percenta year per capita), but with macroeconomic instability.During this period, Mexico suffereddouble-digit
and largedevaluationsin both1976 and 1981. Mexico's govannualinflation
incurredlargefiscaldeficits,and public sectorexternaldebtsoared.
ernment
The two initialstages were characterizedby a semi-closedeconomywith
barriers.Duringthefirststage,inequalityremained
hightariffand nontariff
whereas
duringthesecondstagetherewas a rapidreduction
relativelystable,
in incomeinequalityin Mexico (Szekely 2005).
The thirdstage(1982-94) was one of structural
adjustmentand important
economic reforms.During this period Mexico went througha process of
macroeconomicadjustmentthat led to a radical change in its economic
therewas an
model:thegovernment
drasticallyreducedpublicexpenditures;
sector
of
debt;
large-scaleprivatizaforeign
renegotiation public
important
tiontookplace; and, in 1985, in themidstof an unexpectedcollapse in the
openedup its economyby
priceof itsmainexport(oil), Mexico unilaterally
barriers.
most
ofitsnontariff
and
its
tariffs
eliminating
reducing
significantly
In the early 1990s, Mexico announcedits intentionof going well beyond
thosereforms
(and lockingthemin) byproposinga freetradeagreementwith
theUnitedStatesand Canada.3The agreementwentintoeffectin 1994 as the
NorthAmericanFree Trade Agreement(NAFTA), establishingthe largest
freetradearea in theworld- and themostasymmetrical.
Duringthisperiod
theMexican economystagnatedin per capita termsand income inequality
theperiod.
increasedsubstantially
throughout
The firstyear of the post-NAFTAperiod was characterizedby a severe
macroeconomiccrisisthatbegan in December 1994. In thatmonth,Mexico
a financialdefault.
experienceda largedevaluationand was close to incurring
The fiscaland macroeconomicadjustmentof 1995 led to a sharpand steep
decline in economicactivityduring1995 (a contractionof 8 percentin per
capitaGDP). Later,from1995 to 2000, thedomesticeconomyrecoveredrelincreasein Mexican exports
ativelyquickly,mainlybecause of an important
to theU.S. market.Between 1995 and 2000, Mexico's percapita GDP grew
at a rateof 4 percenta year.
The firstpost-NAFTA stage was also characterizedby the implementasocial and economicprograms:Progresa(laterknown
tionof two important
as Oportunidades)and Procampo.The firstprogramis a targetedconditional
consideredthe
cash transfer
programthatstartedin 1997 and is currently
was
in
Mexico.
mostimportant
implemented
Progresa
program
antipoverty
issues.
details
onthese
andEsquivel
3. SeeTornell
(1997)formore

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 159
firstin ruralareas, althoughithas includedurbanareas since 2001. The second program,Procampo,is an income-support
programforagricultural
proface
the
transition
froma closed economyto a
ducersdesignedto help them
more open economy. The program,which began in 1994 when NAFTA
wentinto effect,is considereda badly designed programin redistributive
terms(Esquivel, Lustig,and Scott2010). On average,theperiod 1994-2000
was one of mediocreeconomic performance(2 percentgrowtha year),but
it was also the period duringwhich income inequalitystartedto fall. The
mostrecentstage,from2000 to 2006, was one of low growthwithmacroeconomic stability.During those years,Mexico's per capita GDP grew at
only1 percenta year,because itwas negativelyaffectedbytheU.S. recession
of 2000-01. Nevertheless,duringthisperiodincomeinequalitywas reduced
even further.

IncomeInequalityin Mexicosince NAFTA


to clarifywhatmeaBeforeinequalityin Mexico is discussed,it is important
ofincomeis used inthispaper,sincedifsureofinequalityandwhatdefinition
estimatesof inequalitybut
could lead notonlyto different
ferentdefinitions
conclusions.4Most of theresults,however,arerobust
also to slightlydifferent
measuresof inequality.
of incomeand alternative
to alternative
definitions
In this paper the Gini coefficientis used as the preferredmeasure of
This measurenotonly satisfiesall thedesirablepropertiesof an
inequality.5
inequalitymeasure6butalso can be decomposedby incomesource,a feature
here.On theotherhand,inequalityis usuallymeasuredusing
thatis ofinterest
are
totalincomeor currentmonetaryincome.7Both definitions
eithercurrent
income
estimates
used in theinitialestimatesofinequality,butlatermonetary
are theonlyfocusof attention.
Figure2 providesa simpledescriptionof the
inMexico
estimates
a survey
ofGinicoefficient
4. Corbacho
andSchwartz
(2002)include
anddifferent
income
definitions.
fordifferent
periods
ofthedistribution.
wellchanges
attheextremes
doesnotcapture
5. TheGinicoefficient
inthetext
similar
tothosedescribed
ofinequality
thatshowtrends
other
measures
However,
alteronrequest.
SeealsoCampos
areavailable
from
theauthor
(2008)forcomparisons
using
native
measures
ofinequality.
tothePigou-Dalton
transfer
adherence
areas follows:
6. Theseprinciples
symprinciple;
anddecomposability.
ofscale;homogeneity;
metry;
independence
This
nettotal
income.
that
iswidely
usedinMexico:
isa third
definition
ofincome
7. There
Thismeaandin-kind
transfers.
totalincome
butdeducts
definition
is similar
tocurrent
gifts
rates
inMexico.
estimation
ofpoverty
sureistheoneusedintheofficial

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

160

EC0N0MIA,Fall2011

FIGURE 2. Sources
ofCurrent
TotalIncome
inMexico
^

-Labor
Income

-OwnBusinesses
Monetary
-Assets
Income
Income *Remittances
Current
. public
Transfers
-
-Transfers
Total J
pens,ons
^
Income '
Private
Transfers
Non-Monetary
Income
V
Source:
ENIGH
(various
years).

The descriptionof the sources of


componentsof both income definitions.
income
are
later
used
in
a
Gini
monetary
decompositionexercise.All estimatesuse information
fromtheNational Surveyof Household Income and
Expenditure(ENIGH, theacronymin Spanish).Surveysare available forthe
years1984, 1989, 1994, 1996, 1998, 2000, 2002, 2004, 2005, and 2006.
Figure 3 shows the evolutionof the Gini coefficientin Mexico forthe
period 1984-2006, using alternativedefinitionsof income. The figure
clearlyshows theexistenceof an invertedU shape thatpeaks in 1994 in all
cases and thatsteadilydeclines thereafter.
The figurealso shows therapid
increase in inequalitythattook place between 1984 and 1994, which has
been reportedin, among other studies, Bouill6n, Legovini, and Lustig
(2003) and Legovini,Bouill6n, and Lustig (2005). The Gini coefficientfor
currentmonetaryincome droppedfrom0.564 in 1994 to 0.505 in 2006, a
10 percentreduction;thecorrespondingmeasurefortotalincome dropped
from0.537 to 0.494, an 8 percentreduction.These reductionsare similarin
magnitudeto thoserecentlyobservedin Brazil and documentedin Barros
and others(2010). In annualterms,inequalityin Mexico has fallenat a rate
of 0.9 and 0.7 percenta year in the case of currentmonetaryincome and
totalincome,respectively.Althoughthoserates are stillbelow the annual
1 percentrate at which income inequalitydiminishedin Mexico between
1954 and 1984,theyshowa significant
withrespectto the1994
improvement
figures.
results.For example,thefigFigure3 also shows a few otherinteresting
ure shows thatthedistribution
of monetaryincomeis moreunequal (thatis,

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 161
FIGURE 3 . Mexico:
Gini
Coefficients
forAlternative
Income
1984-2006
Definitions,

Source:
Authors
elaboration
based
onENIGH
(various
years).

of totalcurrentincome
it has a higherGini coefficient)
thanthedistribution
income
is less unequallydisis
the
fact
that
(which explainedby
nonmonetary
or
on theotherhand,thefigurealso showsthatbeforeeithertransfers
tributed);
remittances
are included,theGini coefficient
of monetary
incometendsto be
of monetaryincome,thussuggesting
slightlyhigherthantheGini coefficient
ofthesetwofactors,an issue discussedlaterin the
thefinalizingcontribution
paper.
inMexico
TheUrban/Rural
Dimension
ofInequality
thedynamicsof
Previousstudieshave showntherelevanceof understanding
and
urban
in
Mexico
For
rural
inequality
separately. example,Pnuco-Laguette
and Szekely (1996) showed that inequalitywithinurban and rural areas
accountedfor84 percentof totalinequalityin Mexico in 1992, whereasonly
one-sixthoftotalinequalitywas explainedbytherural/urban
gap. Forthatreaon
the
of
in
this
now
focuses
son,
dynamics inequality ruraland urban
paper
is crucialto
areas in Mexico since 1994. As discussedlater,thisdistinction
the contribution
of different
factorsin the recentdownward
understanding
trendin inequalityin thecountry.

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

162

ECONOMIA,Fall2011

FIGURE 4. Urban
Mexico:
Gini
Coefficients
forAlternative
Income
1994-2006
Definitions,

Source:
Author's
elaboration
based
onENIGH
(various
years).
forurbanand
Figures4 and 5 show theevolutionof theGini coefficients
ruralareas in Mexico forthe 1994-2006 period.8The divergencein thepatternsofinequalityby sectoris quitestriking.
On one hand,incomeinequality
inurbanareasinMexico,regardlessoftheincomedefinition
used,has steadily
declinedsince 1994. On theother,incomeinequalityin ruralareas increased
until2000, accordingto thetotalincomedefinition,
or until2002, according
to anyotherincomedefinition.
Afterreachingits peak, incomeinequalityin
ruralareasbasicallyreturned
to its 1994 level. The existenceof sucha differentiatedpatternof incomeinequalityin ruraland urbanareas somehowsugfactorscouldbe affecting
thedynamicsinthosetwosectors
geststhatdifferent
of theMexican economy.This idea is exploredin moredetaillater.
TheDistribution
ofMonetary
inMexico
Income
The distribution
ofmonetary
incomein Mexico is now exploredin moredetail
by lookingat thegrowthincidencecurves(GICs) suggestedby Ravallionand
8. Pleasenotethat
thisisnota rural/urban
income
exercise.
This
inequality
decomposition
refers
within
rural
andurban
areasanddoes
analysis
onlytotheincome
dynamics
inequality
notdiscuss
thecontribution
ofeachsector
inMexico.
tototal
inequality

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 163
FIGURE 5. Rural
Mexico:
Gini
Coefficients
forAlternative
Income
1994-2006
Definitions,

Source:
Author's
elaboration
based
onENIGH
(various
years).

Chen(2003). These curvesshowthepercentchangeinpercapitaincomealong


theentireincomedistribution
betweentwopointsin time.Figure6 showsthe
GIC fortheentire1994-2006 periodat thenational,urban,and rurallevels.
The negativeslope in thefirstgraphclearlyshows whyMexico's income
inequalitydiminishedduringthisperiod:incomeat thebottompartofthedistribution
grewfasterthanincomeat themiddleand thetoppartsof thedistribution.The figurealso showsthedifferent
followedby theurbanand
patterns
ruralincomedistributions
this
In
urban
areas, incomegrowth
during
period.
was prettyflatacross theentiredistribution
exceptforthetop threedeciles,
whichexperiencedsmallerand in some cases even negativeincomegrowth
rates.In thecase of ruralareas,two aspectsare salient:first,
averageincome
in
was
than
urban
areas
effect
that,given the relatively
(an
growth
greater
largerural/urban
gap,reducesinequality),and second,theruralGIC curvealso
has a negativeslope, so thatthebottomhalfof theruralincomedistribution
had higherincomegrowthratesthanthetop segmentof thedistribution.
All
thesefactshave contributed
to thereductionin incomeinequalityin Mexico
thathas takenplace since 1994.
theseresultsalreadysuggestthatthereductionin inequality
Interestingly,
in Mexico between 1994 and 2006 came fromdifferent
sources: in urban

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

164

ECONOMIA,Fall2011

Growth
Incidence
Curves
1994-2006
FIGURE 6. Mexico:
Income,
Using
Monetary

Source:
elaboration
onENIGH
Author's
based
(various
years).

areas, it was the resultof the relative(and forsome even absolute) loss of
incomeat thetop partof thedistribution,
whereasin ruralareas, it was the
in ruralincomesas well as the specificimprovegeneralizedimprovement
mentin the income of the relativelypoor ruralhouseholdsthroughout
this
period.

DecompositionAnalysisofSourcesof IncomeInequalityin Mexico


A decompositionof the Gini coefficient
is conductedbelow forthe years
of different
income
1994, 2000, and 2006 to investigatethe contribution
sourcesto theobservedinequalityof monetary
incomein Mexico.

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 165
Sources
ofMonetary
Income
The maincomponentof monetaryincomein Mexico is laborincome,which
accountedforaround60 percentof all monetaryincomein 2006; thesecondlargestsource of monetaryincome in Mexico is income obtainedfromthe
businessesof self-employed
individuals,whichaccountsforanother20 percentof monetary
income.The restof monetaryincomeproceedsfroma variand remittances.
etyof sources,includingtransfers
Table A1 in theappendixshowsthepercentageof householdsthatreceive
incomefromsourcesotherthanlaborincome.This table shows thedramatic
increasethathas takenplace since 1992 in thepercentageof Mexican households thatreceive some typeof transfer.Althoughless than24 percentof
in thatyear,by
householdsreceiveda public or a privatemonetarytransfer
2006 morethan45 percentof all householdsreportedreceivingpartof their
The singlemostimpormonetaryincomethougha privateor publictransfer.
tantcontributor
to this trendis undoubtedlythe social programProgresa/
which,accordingto 2006 ENIGH data,is receivedby 15 perOportunidades,
cent of Mexican households.9Two otherfactorsaccount for part of the
increase in transfersto Mexican households: first,the ruralprogramProcampo, whichwas intendedto supportruralproducersduringthetransition
to tradeliberalizationin agriculturalproducts,10
and, second, remittances,
whichare now receivedin 7 percentof Mexican households,twice as many
as in 1994. Based on whatis knownaboutthedistributive
effectsof theProand
campo (regressive) Progresa/Oportunidades
(veryprogressive)programs
(Esquivel et al. 2010), it is quite likelythattheycan actuallyaccountfora
greatdeal of theup-and-downdynamicsof incomeinequalityin ruralareas
depictedin figure5.
Methodology
Lermanand Yitzhaki(1985) showedthattheGinicoefficient
fortotalincome
inequality(G) withK incomesourcescan be expressedas
(1)

G = tstGkRk,
k=l

9. Formoredetailsaboutthisprogram,
see Corbacho
andSchwartz
(2002)andLevy
(2006).
10. Formore
details
onProcampo,
seeCorbacho
andSchwartz
(2002).

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

166

ECONOMI
A, Fall2011

whereSkis theshareof sourcek in totalincome,Gkis theGini coefficient


of
theincomesourcek,andRkis theGinicorrelation
betweenthesourceincomek
and totalincome.
This decompositionof theGini coefficient
has a neat and clear-cutinterof incomesourcek to inequalpretationsince it shows thatthecontribution
of threeelements:how important
theincome
itydependson theinteraction
sourceon totalincomeis (Sk); how unequallydistributed
theincomesource
is (Gk); and how correlatedthe income source and the distribution
of total
incomeare ( Rk).
Therefore,an incomesourcethatrepresentsa relativelarge shareof total
incomecould have a largeeffecton inequalityas long as it is unequallydistributed
(thatis, ifithas a relativelyhighGk). However,ifGkis low, thisfactorwill dwarfthecontribution
ofthatincomesource.On theotherhand,ifan
incomesourceis veryunequallydistributed
butis nothighlycorrelatedwith
totalincome(as in thecase of well-targeted
transfer
programs),thenthecontribution
of such a sourcecould in factbecome negative.
Stark,Taylor,and Yitzhaki(1986) showedthatwiththistypeof decompositionone can estimatethe effectof a small percentagechange (n) in a
given income source on totalinequality(holdingall otherincome sources
constant)throughthefollowingexpression:
(2)

||

St(Gt*t-G)

or,alternatively,

(3)

_Ak=^L_S^

This expressionmeansthatthepercentchangein inequalityresultingfroma


marginalpercentagechangein incomesourcek is equal to theinitialshareof
incomesourcek on totalincomeinequalityminustheinitialshareof income
sourcek.
GiniDecomposition
Results
Now themonetary
incomeGini coefficients
forMexico are decomposedfollowingtheapproachjust describedand usingtheincome sourcesdescribed
in figure2 and tableAl. For simplicityof exposition,insteadof applyingthe

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 167
methodologyto thewhole periodunderanalysis,itis appliedonlyto thesurveys of 1994, 2000, and 2006. The descogini Stata commandpresentedin
Lopez-Feldman(2006) is used in thedecompositionexercise.
The marginaleffectsof thedecompositionexerciseare shownin figure7.
Results are unequivocal: at the national level, thereare threeinequalitysourcesofincome.Amongthefirst
and threeinequality-reducing
augmenting
groupare pensions,incomefromown businesses,and incomefromproperty
rents.Amongthesecondgroupare incomelabor(at least since 2000), remitIn thelast two cases, themarginalnegativeeffectson
tances,and transfers.
theperiod.
increasedthroughout
have
coefficient
theGini
incomesourcesin
Figure7 also showsthemarginaleffectof thedifferent
income
urbanand ruralareas.The signofthemarginaleffectsofthedifferent
There
level.
the
national
at
as
that
observed
the
same
is
components basically
relative
in termsof the
differences
are,however,some important
importance
oftheimpactof some sources.For example,laborincomeis a veryimportant
forcein urbanareas but not in theruralsector(there,it
inequality-reducing
are a very
even augmentedinequalityin 2006). On theotherhand,transfers
factorin ruralareas but not as importantin
importantinequality-reducing
do not seem to have a large negaurbanones. Finally,notethatremittances
tivemarginaleffecton inequalityin anyspecificsector,althoughtheyarerelevantat thenationallevel. This apparentparadoxis explainedby thefactthat
incomeis close
withruralmonetary,
whiletheGini correlationof remittances
income
to 50 percent,theyhave a muchlowerGinicorrelationwithmonetary
have an effectat thenational
at thenationallevel. In thatsense,remittances
in thebottomhalfof thenational
level because theyare heavilyconcentrated
workas an inequality-reducing
remittances
incomedistribution.11
Therefore,
income gap and not throughthe
source of income throughthe rural/urban
incomedistribution.
sector-specific
WhyLaborIncomeHas Become an EqualizingIncome Force
The resultsof thedecompositionexercisesuggestthatone of themostimportantequalizingforcesin recentyearsinMexico has beentheevolutionoflabor
income,bothin urbanareas and in thecountryas a whole. In fact,thereducaccounts
of laborincometo theGini coefficient
tionin thetotalcontribution
the
for almost all of the observedreductionin this coefficientthroughout
andHuerta-Pineda
onthisissuesee,forexample,
11. Formore
details
(2007).
Esquivel

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

FIGURE 7. Marginal
Effect
onGini
Coefficient
Source:
Overall,
Urban,
byIncome
andRural
Mexico

Source:
Author's
elaboration
onENIGH
based
(various
years).

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 169
thenatureof thechangein the
1994-2006 period.Therefore,
understanding
effectof laborincomeon inequality,whichwentfrombeingpositivein 1994
to becomingnegativein 2000 and 2006 (see figure7), is crucial to understandingthewholedynamicsof incomeinequalityin Mexico since 1994.
To beginwith,notethatlaborincomeis basicallytheproductof multiplying hourlywages by numberof hoursworked.That being thecase, leaving
aside changesin thenumberof hoursworkedalong theincomedistribution
(whichcould have occurredbutprobablynotnecessarilyin themagnitudeor
directionthatcould actuallyexplaintheobservedchangesin incomeinequality),the only otherchannelthroughwhichlabor income can affectincome
mostofthechangesin
changesin wage rates.Therefore,
inequalityis through
thistypeof inequalitysomehowmustbe the outcomeof changes in wage
circumstance,since a link
inequality.In some sense,thisis a veryfortunate
can thenbe establishedbetweenthisdiscussionon incomeinequalityand the
on wage inequalityin Mexico thathas been writtenas partof the
literature
betweentradeand wages.12
debateon therelationship
Let us look firstat the evolutionof wage inequalityin Mexico in recent
of wage inequalitygivenby theratioof
yearsby usinga standarddefinition
workers.Thisratio
workerstothoseofproduction
thewages ofnonproduction
is also (grossly)definedas theskilled/unskilled
wage ratio,whereinnonproductionworkersareconsidereda proxyforskilledlaborandproductionworkers a proxyforunskilledlabor.13
Figure8 showstheevolutionof thismeasureof wage inequalityin Mexifortheperiodfrom1984 through2007. The data forthisgraph
can industry
come fromtheEncuestaIndustrialMensual (EIM) [MonthlyIndustrialSurvey],whichhas monthlyand annualdata on totalwages paid and totalhours
workers.This figworkedin industry
by bothproductionand nonproduction
ure is an updated version of similarversions published in, for example,
Esquivel and Rodriguez-Lopez(2003) and Chiquiar(2008).
The patternof wage inequalityin Mexico's industrialsectorin figure8 is
similarto theevolutionof inequalityunderthevariousdefinitions
remarkably
of incomethatwere shownbefore.This figureshows a continuousupward
increasein wage inequalitysince 1984 thatlasted untilthe mid-1990s,folbetweenthisgraph
lowed by a steadydecline since then.A slightdifference
inGoldberg
andPavcnik
casethat
totheMexican
references
12. Seetheabundant
appear
andinequality.
onglobalization
(2007),a survey
whoare
workers
areproduction
sincethere
a grosssimplification,
13. Thisis,ofcourse,
unskilled.
whoarerelatively
workers
skilled
andnonproduction
highly

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

170

EC0N0MIA,Fall2011

19M~2007
toUnskilled
Industrial
FIGURE 8. Ratio
ofSkilled
Wages,

Author's
elaboration
based
onEIM
Source:
(various
years).
and theincomeinequalityindicators,however,is thatthismeasureof wage
of inequalitypeak
inequalitypeaks in 1996, whereasall theotherdefinitions
14
is thatwage inequalityin 2006, unlikethe
around1994. A seconddifference
to itsmid-1980slevel. That
incomeinequalitymeasures,had notyetreturned
those
associated
withwage inequality
that
some
elements
besides
suggests
to the reductionof income inequalityin Mexico (such as
are contributing
remittances
and transfers,
as discussedabove).
Let us now take a moredetailedlook at theevolutionof wage inequality
inMexico's industry
inrecentyears.Figure9 showstheskilled/unskilled
wage
sectorat two difindustriesin Mexico's manufacturing
ratioforforty-eight
ferentpointsin time.The top image comparesthe observedwage ratioin
1988 (x axis) withthatof 1994 (y axis); thebottomone showstheequivalent
ratiofortheyears 1994 and 2007. Both figuresincludea 45-degreeline as a
reference.The top figureshows thatthe increasein the wage gap between
skilled and unskilledworkersthatoccurredbefore 1994 was generalized
In fact,thewage ratioincreasedin
acrosstheentiremanufacturing
industry.
14. Seealsothediscussion
onthisissueinRobertson
(2008).
(2007)andCampos

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 171
FIGURE 9. Ratio
ofSkilled
toUnskilled
Wages,
byIndustry

Source:
Author's
elaboration
based
onEIM
(various
years).

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

172

ECONOMI
A, Fall2011

oftheforty-eight
branches.Between1994 and 2007,
forty-six
manufacturing
however,thepatternof the skilled/unskilled
wage ratioin Mexico's manulooks
somewhat
different
and
moreheterogeneousthanin
facturing
industry
thepreviousperiod:mostindustries
now showa slightlydecliningwage ratio
betweenthesetwoyears,buttherealso are a fewbranchesin whichthewage
ratiois now eitherthesame or slightlyabove its 1994 level.
On theotherhand,dataon theevolutionof theskilled/unskilled
wage ratio
at thestatelevel show also a clearlydecliningtrendin almosteverystatein
Mexico sincethemid-1990s(Esquivel2008). In summary,
since1996therehas
been an important
reductionin wage inequalityin Mexico. This reductionhas
takenplace notonlyat theindustry-wide
level butalso in mostmanufacturing
branchesand across the countryin manyregionsand states.Consequently,
a good explanationof labor incomeinequality(and of wage inequality)has
to be able to explainnotonlytherapidincreasein wage inequalitybetween
1984 and 1996 butalso thereduction
in wage inequalitythathas beenobserved
since 1996.

Explainingthe EvolutionofWage Inequalityin Mexico


The rapidincreasein wage inequalitythatoccurredin Mexico between1984
and 1994 or 1996 has been widelydocumentedand studied.15
An interesting
aspectof thistrendis thatitsbeginningcoincidedwiththeunilateralopening
oftheMexicaneconomythatstartedpreciselyin themid-1980s.The increase
in Mexico's wage inequalitytherefore
would be somewhatunexpected,consideringthatMexico has a relativeabundanceofunskilledlabor(at leastfrom
theperspectiveof its maintradepartner,
theUnitedStates)and thatstandard
tradetheorieswould have predictedexactlythe oppositepattern(thatis, a
reductionin theskilled/unskilled
wage ratio;see Craggand Epelbaum 1996).
As a consequence,severalpossiblechannels(mostofthemlinkedto theopening of the economyin the mid-1980s)have been suggestedto explain this
apparentparadox.
The explanationsthathave been proposedto explain the post-openness
increaseinMexico's incomeinequalitycan be grosslydividedintotwogroups:
in thefirst,
theexplanationsemphasizefactorsaffecting
thebottompartofthe
15. See,forexample,
(2005),
EsquivelandRodriguez
L6pez(2003),AirolaandJuhn
Robertson
andMontes-Rojas
(2007),Acosta
(2008),Chiquiar
(2008),Verhoogen
(2008),and
thereferences
citedtherein.

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 173
incomedistribution
(thatis, thesegmentthatforthemostpartcomprisesless
skilledand less experiencedworkers);thesecondemphasizesfactorsaffecting
In thefirstgroup,forexample,are theories
theupperpartof thedistribution.
emphasizingthereductionin realminimumwages (Fairris,Popli,and Zepeda
2008) as well as theoriessuggestingthatthe mid-1980sreductionin tariffs
affectedindustries
thatemployedmostlylow-skilledworkdisproportionately
ers (Hanson and Harrison1999). In the second group,some explanations
emphasizetheroleof an increasein thedemandforskilledworkersassociated
withthe presenceof foreigninvestment
(Feenstraand Hanson 1997); with
skill-biasedtechnologicalchange (Cragg and Epelbaum 1996; Esquivel and
Rodriguez-Lopez2003); and with a quality-upgrading
process due to an
increase in exports(Verhoogen2008). Otherexplanationshave suggested
thateducationinequalitycould have also playeda role(Lopez-Acevedo 2006)
or thatthese trendscould indicate only short-runeffects(Canonero and
Werner2002).
On theotherhand,thepost-1996 reductionin wage inequalityin Mexico
has been much less studied. So far,only Robertson(2007) and Campos
(2008) have analyzedthistrend.While Campos favorsan explanationbased
workers
on supplyfactors,RobertsonsuggeststhatMexico's manufacturing
are now complementsof ratherthansubstitutesforU.S. workersand that
therehas been an important
expansionof assemblyactivitiesin Mexico that
has increasedthedemandforless skilledworkers.
Of course,manyoftheproposedexplanationsforthepre-NAFTAincrease
in wage inequalityin Mexico are notmutuallyexclusive,and theycould in
factbe at least partiallycorrect.However,it is also truethatmostof them
cannotexplain the subsequentreductionin wage inequalitythathas been
observed since 1996. Thereforethese explanationsare eitherincorrector
incomplete,since therecould be manyunderlyingforcesactingin different
directions.That is why Robertson(2007) noted thatthe patternof wage
inequalityin Mexico is puzzlingbecause no singletheorycould explainthe
evolutionof wage inequalitybeforeand afterNAFTA.16
Althoughit is not the objective of this paper to identifyor to establish
whichexplanation(ifany) is correct,at least some ofthemcould be ruledout
For
forsucha pattern.
sometentative
theoretical
16. Thereare,however,
explanations
under
certain
evenifthestandard
Atolia(2007)suggested
circumstances,
that,
preexample,
inthelongrun,
there
model
works
aspredicted
diction
from
a Heckscher-Ohlin-Samuelson
may
that
leadtoa different
outcome
effects
oftrade
liberalization
besomeshort-run
(ortransitory)
ofsomesectors
inthecontraction
andexpansion
because
oftwofactors:
anasymmetry
first,
inproduction.
thecapital-skill
and,second,
complementarity

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

174

ECONOMI
A, Fall2011

by lookingat some wage dataprovidedby Campos (2008). The nexttwofigures show themean log wage of male workersin Mexico forselectedyears
combinationsof educationand yearsof experience.Workand fordifferent
ersareclassifiedaccordingto thelevel ofeducationachieved(less thatlowerand college education)and to
upper-secondary,
secondary,lower-secondary,
thenumberof yearsof workexperience(less or morethantwentyyearsof
experience).
The upperpartof figure10 showsdata fortheyears1989, 1994, and 1996
for 1996 and 2006. The firstfigure
and the lower partshows information
result:between1989 and 1994, mostof thechangesin
shows an interesting
in Mexico occurredin theuppertailof thedistribution.
thewage distribution
Thatis, theincreasein wage inequalityin thoseyearscannotbe explainedby
a reductionin thewages oflow-skilledor inexperiencedworkers;instead,the
increasecan be explainedonlyby an increasein thewages of highlyskilled
orhighlyexperiencedworkers.This resultbasicallyrulesoutanyexplanation
such as those
based on changes in the lower tail of the wage distribution,
based on a fallingreal minimumwage or on a biased opennessof unskilled
labor-intensiveindustries.This figurealso shows the widespreadnegative
effectsof thefinancialcrisisof 1994-95, whichreduced,almostproportionally,therealwages ofall typesofworkersin Mexico between1994 and 1996.
in Mexico for
The bottompartof figure10 shows the wage distribution
1996 and 2006. Unlike figure9, thisone shows thatmostof thechangesin
thewage distribution
tookplace in thelowertail.Thatis, workerswithlower
levels of educationand/orfeweryears of workexperiencehad the largest
increasesin theiraverage wages, and thatexplains the reductionin wage
inequalitythathas been observedsince 1996. This also suggeststhatanyconvincingstoryofthepost-NAFTAreductionin wage inequalityhas to explain
theincreasein thewages of low-skilled/less
experiencedworkersratherthan
thereductionin thewages of high-skilled/more
experiencedworkers.
The previousresultsconfirmtheintuitionthatthereis no singleexplanationfortheevolutionof wage inequalityin Mexico since 1984. Indeed,the
factthatthe 1984-94 increasein wage inequalityis associatedwithchanges
in theuppertailof thedistribution
whilethepost-NAFTAreductionin wage
inequalityis associatedmostlywithchangesin thebottomtail suggeststhat
thereare at leasttwoleadingforcesat play.In thefirstcase, as discussed,the
onlyexplanationsthatseemtobe compatiblewiththeobservedtrendarethose
suggestingthe presenceof skill-biasedtechnologicalchange,eitherexogenous change(Cragg and Epelbaum 1996 and Esquivel and Rodriguez-Lopez
2003) or endogenouschange resultingfromthe presenceof multinational

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 175

andExperience,
Select
Years
FIGURE 10. MeanLogWageofMaleWorkers
byEducation

based
onCampos
Source:
Author's
elaboration
(2008).

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

176

ECONOMI
A, Fall2011

firms(Feenstraand Hanson 1997) in Mexico and/orby theupgradingof the


qualityof exportingfirms(Verhoogen2008).
For thepost-NAFTAperiodthereare at least threepossible explanations,
two of whichhave alreadybeen mentionedand are notmutuallyexclusive:
an increasein the supplyof relativelyskilled workers(Campos 2008) and
an increasein the demandforunskilledlabor resultingfroman expansion
sector(Robertson2007).
of assemblyactivitiesin Mexico's manufacturing
Eitherof thetwoeffectscould explainthereductionin theskilled-wagepremiumobservedin thedata. A thirdexplanationthatis also compatiblewith
theprevioustwois thatofa standardHeckscher-Ohlin
effectin a countrysuch
as Mexico in whichunskilledlaboris abundant(Chiquiar2008). This effect
could be thelate outcomeof tradeliberalizationsuggestedby Canoneroand
Werner(2002) and alreadymodeled by Atolia (2007) or, alternatively,
an
not
show
in
data
effect
that
did
the
before
due
to
the
underlying
up
presence
of a stronger
force,suchas a skill-biasedtechnologicalchangeas previously
hypothesizedby Esquivel and Rodrfguez-L6pez(2003).
A much more detailed and rigorousanalysis is needed to discriminate
amongthesealternative
hypotheses.However,it is possibleto move forward
byanalyzingwhethersomeofthesehypothesesareborneoutbythedata.Figure 11 showsthecompositionofMexico's workforce
between1989 and 2006
to
the
levels
of
education
and
defined
above. This comaccording
experience
reflects
the
both
interaction
of
positionobviously
supplyand demandfactors.
In general,thefigureshows thatthroughout
theperiodtherewas a large
reductionin theshareoftheleastskilledand leastexperiencedworkers(those
withless thansecondaryeducation)and an increasein thesharesof theother
typesof workers.The most dramaticchanges,however,took place in the
share of workerswith less than secondaryeducation.In fact,this group,
whichaccountedforalmost55 percentof theworkforce
in 1989,represented
about
one-third
of
the
workforce
a
reduction
of about20 peronly
by 2006,
centagepointsin a seventeen-year
span.Thatreductionwas compensatedfor
by increasesin the shares of most of the othergroupsof workers.These
trends,whichwere alreadypresentbetween 1989 and 1994, acceleratedin
thepost-NAFTAperiod.
These resultstherefore
suggestthatat least partof therelativeincreasein
thewages of thelow-skilled/less
experiencedworkersin Mexico is associated withthechangein thecompositionof theworkforceand, in particular,
witha reductionin thenumberof unskilledworkersratherthanan increase
in thesupplyof skilledworkers.Of course,thisresultis notat all incompatible withthehypothesisthatthedemandforunskilledworkersincreased,as

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 177
FIGURE 11. Workforce
andExperience,
1989-2006
Composition
byEducation

onCampos
Source:
Author's
elaboration
based
(2008).
suggestedin Robertson(2007), but,by itself,the resultcannotexplain the
simultaneousincreasein therelativewages and reductionin theparticipation
of theseworkersin Mexico's totalworkforce.
Finally,figure12 shows some resultsthatare compatiblewiththe view
thatemphasizes the role of the compositionof the labor force.The graph
shows on thex-axis thechange between 1996 and 2006 in the shareof the
eight different
groups of workersaccording to theirlevels of education
and experienceas definedabove. Participationin Mexico's workforcehas
declined in threegroups,which correspondto the least educated and less
experiencedworkers.The y-axisindicatestheaveragechangein thelog wage
of male and femaleworkersthatbelong to each of thegroups.As expected,
the groupswhose shareshave diminishedin the past decade are thosethat
have had thelargestincreasein wages. Notice thattheincreasesin thewages
- and in some cases even close to
of theseworkersare close to 20 percent
30 percent throughout
theten-yearperiod.On theotherhand,thecategories
of workerswhose sharesin Mexico's workforcehave increased(the more
educated/more
experiencedworkers)have tendedto have eitherstagnantor

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

178

ECONOMIA,Fall2011

inShare
FIGURE 12. Change
ofTotal
Workers
andExperience
versus
change
byEducation
inLogWage
1996-2006
Gender,
by

Source:
Author's
elaboration
based
onCampos
(2008).
even decreasingwages since 1996. This graphthensupportsthehypothesis
thatthechangein thecompositionofMexico's workforce
is theleadingforce
in thereductionin wage and laborincomeinequalityin Mexico in thepostNAFTA period.
Summaryand Conclusions
This paper reviewsthepatternof incomeinequalityin Mexico since 1994,
when NAFTA went into effect.Using information
fromnationallyrepresentativehousehold surveys,it shows that there has been an important
reductionin income inequalitysince 1994 and thatthis trendhas almost
reversedincomeinequalityto thelevels thatwereobservedbeforetherapid
increasein inequalitythattookplace between1984 and 1994.
As shown by a Gini decompositionanalysis by income source, labor
income, remittances,and public transfers(mainly throughthe Progresa/

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 179
role in thisequalizing
Oportunidadesprogram)have all played an important
has become a very
In
the
shows
that
labor
income
process. particular, paper
in
in
while
force
urban
areas
Mexico,
public transfers
important
equalizing
have been especially importantin reducinginequalityin the ruralsector.
Remittances,on the otherhand, have been a nationalinequality-reducing
sourceof incomesince 1994.
The paperalso providessome evidencesuggestingthattheforcesthatled
to a sharpincreasein wage inequalityacross all industriesin Mexico during
the 1980s and early 1990s are no longeroperating.In fact,a generalized
reductionin wage inequalityacross industriesand regionsin Mexico is now
observed,suggestingthegrowingrelevanceof otherelementsin thistrend.
In general,I believe that Mexico is now beginningto experiencethe
and of tradeffectsof havinga moreeducatedworkforce
inequality-reducing
to have
This
effect
seems
countries.
equalizing
ing withmoreskill-abundant
been postponedby a skill-biasedtechnologicalchange (eitherexogenousor
endogenous)or by an endogenoustechnologicalupgradingthat,in any case,
now seems to have ended.This factand an ambitiousand widespreadsocial
programfocusedon poor ruralhouseholdsseem to be themain explanatory
factorsin thesharpreductionin inequalitythathas been observedin Mexico
in recentyears.

Appendix
Income
from
Sources
Other
thanLabor
ThatReceive
Income
TABLE A1 . Households
Percent
oftotal
households
Source
ofincome

1992 1994 1996 1998 2000 2002 2004 2005 2006

43.8 42.7 43.3


Own
business
4.4 3.5 3.7
rent
Property
25.9 15.2 22.9
Financial
income
23.5 23.8 29.0
Transfer
3.7 3.4 5.3
Remittance
8.8 8.1 8.2
Pension
Public
and
transfers13.6 14.5 18.7
private
...
1.2 4.6
Procampo
n.a. n.a. n.a.
Progresa/Oportunidades
Source:
Author's
estimates
based
onENIGH
(various
years).

43.2
3.5
19.1
31.2
5.3
9.5
19.6
2.8
n.a.

41.1
3.1
19.1
34.0
5.3
10.0
23.1
2.6
n.a.

41.9
4.0
19.3
38.6
5.7
10.0
28.4
5.6
12.3

38.1
4.7
20.2
42.0
5.6
11.8
31.2
4.6
13.4

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

39.1
4.0
18.1
41.3
6.0
11.4
30.5
3.5
13.5

42.0
4.6
23.5
45.5
7.0
11.9
34.6
4.0
14.8

Comment

GuillermoCruces: The paperbyGerardoEsquivelinthiseditionofEconomia


presentsa thoroughaccountof inequalitytrendsin Mexico since the mida fundamental
extensionofthediscussion
1990s.Paperssuchas thisconstitute
of aggregateregionaltrendssuch as thatprovidedby Gasparini,Cruces,and
to discussin depththefacTornarolli(201 1), and theypresentan opportunity
the
level.Identifying
torsunderlying
theevolutionof inequalityat thecountry
is never
causes of theevolutionof a nationalincomedistribution
fundamental
an easy feat,and theauthormustbe laudedforprovidinga coherentand compact discussionof an eventfulperiodin Mexico thatincludedmajortransformationswithpotentiallylarge effectson theincome distribution,
including
theopeningof the
structural
macroeconomiccrises,market-oriented
reforms,
withtheworld's
economyto international
capitalflows,a freetradeagreement
increased
globalization,a political
largesteconomy,technologicalchange,
and even an armedindigenousuprising.
transition,
This commentexaminessome evidenceon distributional
changesin terms
of povertyreductionto complementthe article's main focus on inequality
and thendiscussestheauthor'sconclusionsin termsof a structural
changein
ecoinequalitytrendsin Mexico in lightof the effectsof the international
nomiccrisisthatstartedin 2007. Finally,itsuggestssome avenuesforfurther
research.
The firstpointof thiscommentconcernstheevolutionof povertyoverthe
periodof timecoveredby thearticle.Nationalaggregatepovertyratesforthe
US$2.50 and US$4.00 purchasingpowerparity(PPP) international
poverty
linesarepresentedin figure1 fortheperiod1989-2008 and also fortheurban
and ruralpopulationsfortheUS$2.50 povertyline.This figureprovidesinforthethorough
mationon a dimensionofdistributional
changethatcomplements
discussionof theevolutionof inequalityin thepaper.The figureindicatesa
iswith
andCONICET.
Guillermo
Cruces
CEDLAS-UNLP
180

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 181
FIGURE 1. Poverty
RateforNational
andRural/Urban
1989-2008
Mexico,
Populations,

Source:
CEDLAS
and
World
Bank
Based
onMexico's
INEGI
biannual
Encuesta
Nacional
deIngresos
delos
(2011).
survey,
yGastos
(ENIGH).
Hogares
moderatefall in povertyin the early 1990s, followedby a large increase
between1994 and 1996 thatcan be attributed
mainlyto themacroeconomic
crisisof 1995. From 1996 until2006, thereis a continuousfallin thepoverty
ratespresentedin the figure.In fact,thisevidenceindicatesthatthenotable
resultsfromfigure4 in thepaper,whichshow substantial
growthin incomeat
thebottomofthedistribution
in theperiod1994-2006,could be even largerif
computedusingthepost-crisis
year 1996 as a basis.
This discussionillustrates
a further
dimensionofthedistributional
dynamics documentedin figure1 of thepaper:thereductionin inequalityin Mexico
was accompaniedby a substantialfallin theproportion
of thepopulationliving underdifferent
povertylines and reflectsthe increase in standardsof
livingand thepro-poornatureofthegrowthprocessin theMexican economy
over theperiodunderstudy.At the same time,thisevidence also indicates
thatthereis stillample room forimprovement:about 15 and 30 percentof
the populationstill lived underthe US$2.50 and US$4.00 povertylines,
in 2008, andmajorregionalinequalitiesareevidentin thepoverty
respectively,

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

182

ECONOMIA,Fall2011

andYearly
GDPGrowth
Rate1
FIGURE 2. Labor
Income
Trend
Index
Poverty

who
Source:
Income
Trend
Index
indicates
the
of
individuals
cannot
cover
the
cost
of
abask
C0NEVAL(2011).1heLabor
proportion
Poverty
food
basket
with
their
labor
income.
Itisbased
onMexico's
INEGI
the
Encuesta
National
deOcupaddn
(ENOE).
quarterly
survey,
yEmpleo
= 1.
a. Fourth
with
tofourth
of
the
base
fourth
of
2005
quarter
respect
quarter
previous
year;
quarter

levels in ruralareas,whichare abouttwicethenationalaggregate.While the


differentials,
paper discusses urban-rural
providingmore evidence on their
natureand discussingtheirconsequences in termsof policy are certainly
worthwhile
directionsforfurther
research.Finally,figure1 reveals a small
butsignificant
increasein povertybetween2006 and 2008; theimplications
of thischangeand itsunderlying
factorsare discussednext.
The secondpointof thiscommentrefersto thesustainability
of thedistributionalchangeprocessdocumentedin thepaperand to itsvulnerability
with
to
shocks.
The
2010
national
household
survey(ENIGH)
respect aggregate
was notavailable at thetimethisissue wentto press,implyingthatno aggregate povertyratescomparableto thosein figure1 were available. Figure2
measure,theLabor IncomePovertyTrendIndexconpresentsan alternative
- structed the
by
Consejo Nacional de Evaluaci6n de la Polfticade Desarrollo
Social [National Council for Evaluation of Social DevelopmentPolicy]
(CONEVAL) and based on INEGI' s quarterlysurvey,theEncuestaNacional

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 183
de Ocupacion y Empleo [NationalSurveyof Occupationand Employment]
(ENOE). The index,whichindicatestheproportionof individualswho cannot cover the cost of a basic food basketwiththeirlabor income,has been
normalizedto 1 forthe fourthquarterof 2005. The evidence in the figure
illustratesa seriesof factors.First,thesizable growthin GDP in 2005, 2006,
and 2007 (between3 and 4 percenteach year) did nottranslateintosubstantial reductionsin aggregatepoverty.In fact,therewas a relativelylarge
to the
increasein urbanpovertyratesfor2005-07, whichcan be attributed
increasein thepriceof food relatedto an upwardtrendin commodityprices
and growthin theglobal economyduringthatperiod.
evidence on the increasein
Second, the figurenot only providesfurther
illustratestheverylarge
2006
and
2008
but
also
between
aggregatepoverty
economic
of
the
international
distributional
crisis,withthe poverty
impact
in the2008-10 period.Thatincreasewas seen even for
rateincreasingfurther
theruralpopulation,whichwas relativelysparedfromthe2005-07 upward
trend.The large increasein povertyfor2008-10 is all the more worrying
crisis had only a relativelymodest
when consideringthatthe international
effect(at least by Latin Americanstandards)on domesticgrowthrates,with
reductionsof about 1 and 2 percentof GDP for2008 and 2009, respectively.
Moreover,the reversalin thistrendbetween2009 and 2010, witha yearly
reductionin
growthrateof about4.5 percent,did nottranslateintoa further
poverty:the Labor Income PovertyTrend Index increased (althoughonly
slightly)overthatperiod.
The evidence in figure1 points toward a more fundamentalquestion
relatedto thenatureandthelimitsoftheprocessofdistributional
changedocumentedby theEsquivel paper fortheperiod 1994-2006. While inequality
over theperiod,thereversalin growthtrends
and povertyfell substantially
resultedin a large increase in aggregatepovertylevels and, apparently,a
change in the mechanismlinkinggrowthwith povertyreduction,as witnessedby thestagnantpovertylevels of 2010. Moreover,theSEDLAC database (CEDLAS and World Bank 2011), on which figure1 in the paper is
between2006 and
based, indicatesan increaseof 0.7 in theGini coefficient
at standardlevels).
2008 (althoughthechangeis not statisticallysignificant
to accomplishat
harder
terms
are
in
distributional
Whilefurther
always
gains
economic
lowerpovertyand inequalitylevels,theimpactof theinternational
overthe
to adjustthepolicies implemented
crisiscould providea framework
previousperiod,especiallyin termsof the breadth,depth,and modalityof
of
safetynetsand also in termsof theimplicationsof theproductivestructure
theeconomyand its linkswithmajoreconomies.

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

184

ECONOMI
A, Fall2011

ofSkilled
andUnskilled
intheSkilled
andRelative
TABLE 1 . Changes
Supply
WagePremium
andRelative
1989-2008
Workers
Mexico,
Demand,
Relative
Relative
Relative
demand
Relative
demand
demand
Wage
Period
(o=4)
premium
supply
(a=2)
(c=3)
7.2
9
10.8
1989-2000
1.8
3.6
- 92
-2.8
2.2
-3.5
-6.3
2000-08
and
onmicrodata
from
household
Source:
Cruces,
Galiani,
(201
1),based
Gasparini
surveys.
inaCES
of
suba. Under
of
substitution
the
of
substitution
alternative
(constant
assumptions;
elasticity
elasticity
oHrepresents
elasticity
=workers
xannual
function.
See
Katz
and
for
details.
100
Skilled
workers
with
some
more
stitution)
(1992)
production
Murphy
log
changes.
=workers
unskilled
workers
with
education
orless.
education;
secondary
tertiary
Thiscomment'sfinalpointis relatedto thediscussionin thepaperofskilled
The
and theirimpacton theaggregateincomedistribution.
wage differentials
in
some
evidence
the
increase
the
skilled
paperprovides
corroborating
wage
premiumovertheearly1990s and an apparentreversalof thistrend,withthe
authorpointingouttheimportance
of educationalupgradingand tradingwith
moreskill-abundant
countries
as underlying
factors.The multiplepiecesofeviin a Katz and Murphy(1992)
dence presentedin thepapercan be integrated
as done foran earlierperiodin Mexico in MontesRojas (2006)
framework,
and Manacorda,S&nchez-Pdramo,
done
and Schady (2010) and as currently
forseveralcountriesin theregionin Cruces,Galiani,and Gasparini(2011).
Table 1, reproducedfromthe latterstudy,depictsthe change in the skilled
wage premiumand therelativesupplyof skilledto unskilledworkers,as well
as themagnitudeof demandfactorsimpliedby thesechangesunderdifferent
fortheelasticityof substitution
betweenthetwofactors,as in the
assumptions
Katz and Murphy(1992) analysisoftheUnitedStates.The tableindicatesthat
theskilledwage premiumincreasedsubstantially
in Mexico in thelate 1980s
and early1990s and thatitbeganto fallin themid-1990s,earlierthanin most
otherLatinAmericancountries(Gasparini,Cruces,and Tornarolli2011). The
tablepresentsonlythe1989-2008 change,whichindicatesan overallincrease
in theskilledwage premiumeven in thecontextof an increasein therelative
supplyof skilledworkers.However,theevidenceforthe2000-08 periodindiwithan overallnegative
catesan accelerationof thefallin thisskillpremium,
change. The patternfor the 1989-2000 period is consistentwith a strong
increase in the relativedemand for skilled labor, while the evidence for
2000-08 impliesa substantial
reversalin theeffectof demandfactors.
The conclusionof thepapersuggestsimplicitlya permanent
changein the
forces
trends
and
other
distributional
changes.
underlying
drivinginequality
However, this briefdiscussion points out potentialvulnerabilitiesin the

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 185
natureof these changes. While innovativesocial policy certainlyplayed a
substantialrole in thereductionin povertyover theperiodunderstudyand
changes in labor demand patternsreduced inequality,the ever-evolving
natureof theworldeconomy,to whichMexico now seems fullyintegrated,
implies the need to constantlyrevise policy initiativesand the nature,
The paperdoes
effects,and stabilityof theeconomy's productivestructure.
a verygood job of documentingthe trendsand underlyingfactorsforthe
1994-2006 period. A possible next step in the analysis is to determine
economic crisis representsonly a disturbancein
whetherthe international
Mexico's continuingpath towardreductionof inequalityand povertyor
whetherit may have lasting consequences for growthand distributional
dynamics.

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

186

ECONOMIA,Fall2011

References
2008. "TradeReform
andInequality:
Acosta,Pablo,andGabrielV. Montes-Rojas.
inthe1990s."WorldEconomy
TheCase ofMexicoandArgentina
31 (6): 763-80.
andChinhui
Juhn.
2005."WageInequality
inPost-Reform
Mexico."IZA
Airola,Jim,
DiscussionPaper1525.Bonn,Germany:
Institute
fortheStudyofLabor(IZA).
in Latin
and RisingWage Inequality
Atolia,Manoj. 2007. "TradeLiberalization
America:Reconciliation
withHOS Theory."
Journal
Economics
ofInternational
71 (2) (April):467-94.
2010."Markets,
theState,andtheDynamics
ofInequalBarros,
Ricardo,andothers.
inLatinAmerica,
editedbyLuis
ity:Brazil'sCase Study."InDeclining
Inequality
FelipeLopez-CalvaandNoraLustig,pp. 134-74.Brookings.
in
Bouillon,Cesar,AriannaLegovini,and NoraLustig.2003. "RisingInequality
Mexico:HouseholdCharacteristics
andRegionalEffects."
JournalofDevelopmentStudies39 (4): 112-33.
2008. "WhyDid Wage Inequality
Decreasein Mexicoafter
Campos,Raymundo.
NAFTA?"Mimeo.University
ofCalifornia-Berkeley
(October).
andAlejandro
Werner.
2002."SalariosRelativos
Canonero,
Gustavo,
yLiberalization
delComercio
enMexico."Trimestre
Economico
69 (273): 123-42.
CEDLAS andWorldBank.2010and2011. SEDLAC (Socio-Economic
Databasefor
LatinAmericaandtheCaribbean).
Centrode EstudiosDistributivos,
Laborales,
y
Nacionalde La Plata(http://sedlac.econo.unlp.edu.ar).
Sociales,Universidad
and the
Chiquiar,Daniel. 2008. "Globalization,
RegionalWage Differentials,
Theorem:EvidencefromMexico."JournalofInternational
Stolper-Samuelson
74 (1): 70-93.
Economics
CONEVAL (ConsejoNacionalde Evaluationde la Polfticade DesarrolloSocial).
2011. LaborIncomePoverty
TrendIndex.Mexico.
2002. "Mexico:Experience
withExpenditure
Corbacho,Ana,andGerdSchwartz.
Pro-Poor
Policies."Working
International
MonePaperWP/02/12.
Washington:
taryFund(January).
Cragg,M., and MarioEpelbaum.1996. "WhyHas Wage DispersionGrownin
Mexico?Is It theIncidenceof Reforms
or theGrowingDemandforSkills?"
Journal
Economics51 (1): 99-116.
ofDevelopment
in Latin
Cruces,G., S. Galiani,and L. Gasparini.2011. "Wage Skill Premiums
America,1990-2010:Evidenceon Supplyand DemandFactorsforSixteen
Countries."
workin progress.CEDLAS-UNLP and Washington
Unpublished
in SaintLouis.
University
in Chile."
Eberhard,
Juan,andEduardoEngel.2008. "Decreasing
WageInequality
de Chile.
Unpublished
paper.Universidad
EIM ("EncuestaIndustrial
Mexico:InstiMensual").Variousyears.Aguascalientes,
tutoNacionalde Estadistica,
e Informatica.
Geografia,
ENIGH ("EncuestaNacionalde Ingresosy Gastosde los Hogares").Variousyears.
Mexico:Instituto
NacionaldeEstadistica,
e Informatica.
Aguascalientes,
Geografia,

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Gerardo
Esquivel 187
Esquivel,Gerardo.2008. "IncomeInequality
Dynamicsin MexicosinceNAFTA."
RPP-LACWorking
NewYork:UnitedNationsDevelopment
ProPaper03/2008.
Bureau
for
Latin
America
and
the
Caribbean.
gram,Regional
in
andAlejandra
Huerta-Pineda.
2007."Remittances
andPoverty
Esquivel,Gerardo,
Mexico:A Propensity
ScoreMatching
and
Trade
Journal
Integration
Approach."
(July-December).
NoraLustig,andJohnScott.2010."Mexico:A DecadeofFalling
Esquivel,Gerardo,
Market
Forcesor StateAction?"In DecliningInequalityin Latin
Inequality:
America
, editedby Luis Felipe Lopez-Calvaand Nora Lustig,pp. 175-217.
Brookings.
2003. "Technology,
Trade,and
Esquivel,Gerardo,and JoseA. Rodriguez-Lopez.
in Mexicobeforeand afterNAFTA."JournalofDevelopment
WageInequality
Economics
12(2): 543-65.
Fairris,David,GurleenPopli,and EduardoZepeda. 2008. "Minimum
Wage and
inMexico."ReviewofSocialEconomy
LXVI (June):181-208.
WageStructure
and Relative
R. C., and G. Hanson.1997. "ForeignDirectInvestment
Feenstra,
JournalofInternational
EcoWages:EvidencefromMexico'sMaquiladoras."
nomics42 (3-4): 371-93.
Wai-Poi.2007."TradeLibFrancisco
H. G.,PhillipeG. Leite,andMatthew
Ferreira,
inBrazil."PolicyResearch
Flows,andWageInequality
eralization,
Employment
WorldBank.
Working
Paper4108.Washington:
inMotion:TheCase
andGuillermo
Cruces.2010."Distribution
Leonardo,
Gasparini,
In DecliningInequality
of Argentina."
in LatinAmerica
, editedby Luis Felipe
Lopez-CalvaandNoraLustig,pp. 100-33.Brookings.
2009. "Recent
Cruces,and LeopoldoTornarolli.
Leonardo,Guillermo
Gasparini,
inLatinAmerica."
TrendsinIncomeInequality
Working
Paper2009-132.Palma
de Mallorca,Spain:SocietyfortheStudyofIncomeInequality.
. 2011. "RecentTrendsinIncomeInequality
inLatinAmerica."
Economia11
(2): 147-90.
Effects
of GlobalGoldberg,
PinelopiK., andNinaPavcnik.2007. "Distributional
izationinDevelopingCountries."
Journal
XLV (March):
ofEconomicLiterature
39-82.
andAnnHarrison.
1999."TradeLiberalization
andWageInequalHanson,Gordon,
andLaborRelations
Review52 (2): 271-88.
ity."Industrial
1992."ChangesinRelativeWages,1963-1987:Supplyand
Katz,L., andK. Murphy.
DemandFactors."Quarterly
Journal
ofEconomics107(1): 35-78.
CesarBouillon,andNoraLustig.2005."CanEducation
Legovini,Arianna,
Explain
in Mexico?"In TheMicroeconomics
Changesin IncomeInequality
ofIncome
Distribution
F. H. G. Ferreira,
and N.
, editedby F. Bourguignon,
Dynamics
275-312.
World
Bank/Oxford
Press.
Lustig,pp.
Washington:
University
andShlomoYitzhaki.1985."IncomeInequality
Effects
Lerman,
Robert,
byIncome
Source:A NewApproach
andApplication
totheU.S." ReviewofEconomics
and
Statistics
67 (1) (February):
151-56.

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

188

ECONOMI
A, Fall2011

Mexico'sProgresaLevy,Santiago.2006. ProgressagainstPoverty:Sustaining
Program.
Brookings.
Oportunidades
ofEducaGladys.2006. "Mexico:Two DecadesoftheEvolution
Lopez-Acevedo,
World
tionandInequality."
Paper3919.Washington:
PolicyResearchWorking
Bank(May).
Lopez-Calva,Luis Felipe,and Nora Lustig.2010. DecliningInequalityin Latin
America
. Brookings.
andObtaining
2006."Decomposing
Marginal
Inequality
Alejandro.
Lopez-Feldman,
Effects."
StataJournal
6 (1): 106-11.
ofInequality."
theState,andtheDynamics
Lustig,Nora.2007."A NoteonMarkets,
Program,
Unpublished
paper,p. 17. New York:UnitedNationsDevelopment
RegionalBureauforLatinAmericaandtheCaribbean.
to
andN. Schady.2010. "Changesin Returns
M., C. Sanchez-Paramo,
Manacorda,
in LatinAmerica:TheRole ofDemandandSupplyofSkills."IndusEducation
Review63 (2).
trialandLaborRelations
MontesRojas, G. 2006. "SkillPremiain Mexico:Demandand SupplyFactors."
AppliedEconomicLetters13 (14): 917-24.
and
andMiguelSzekely.1996."IncomeDistribution
Humberto,
Panuco-Laguette,
andItsImpact
inMexico."In TheNewEconomic
ModelinLatinAmerica
Poverty
andPoverty
onIncomeDistribution
, editedbyV. Bulmer-Thomas,
pp. 185-221.
Press.
London:Macmillan
Pro-PoorGrowth."
Ecoand ShaohuaChen.2003. "Measuring
Ravallion,Martin,
78 (1): 93-99.
nomicsLetters
2007."TradeandWages:Two PuzzlesfromMexico."World
Robertson,
Raymond.
30
1378-98.
Economy (9):
and
Stark,Oded, J. EdwardTaylor,and ShlomoYitzhaki.1986. "Remittances
722-40.
Economic
Journal
96
(September):
Inequality."
Szekely,Miguel.2005. "Pobrezay Desigualdaden Mexicoentre1950 y 2004."
Economico72 (288) (October-December).
Trimestre
Tornell,Aaron,andGerardoEsquivel.1997."ThePoliticalEconomyof Mexico's
TradeArrangements
versusMultilateral
,
EntryintoNAFTA."In Regionalism
ofChicagoPress.
editedbyT. ItoandA. Krueger.
University
intheMexandWageInequality
Eric.2008."Trade,QualityUpgrading,
Verhoogen,
123(May):489-530.
Sector."Quarterly
Journal
icanManufacturing
ofEconomics

This content downloaded from 200.52.255.1 on Fri, 14 Feb 2014 15:01:28 PM


All use subject to JSTOR Terms and Conditions

Das könnte Ihnen auch gefallen