See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/321275240

The rise and fall of income inequality in Mexico, 1989-2010

Article · January 2014

CITATIONS READS

52 448

3 authors:

Raymundo M. Campos-Vazquez Gerardo Esquivel

The College of Mexico The College of Mexico
110 PUBLICATIONS 1,328 CITATIONS 70 PUBLICATIONS 3,552 CITATIONS

SEE PROFILE SEE PROFILE

Nora Lustig
Tulane University
255 PUBLICATIONS 5,416 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Deconstructing the Decline in Inequality in Latin America View project

Declining Wages for College-Educated Workers in Mexico: Are Younger of Older Cohorts Hurt the Most? View project

All content following this page was uploaded by Raymundo M. Campos-Vazquez on 14 December 2018.

The user has requested enhancement of the downloaded file.

 7

The Rise and Fall of Income Inequality

in Mexico, 1989–2010*
Raymundo Campos-Vazquez, Gerardo Esquivel, and Nora Lustig

7.1 Introduction

During the last twenty years, the evolution of inequality in Mexico has fol-
lowed two distinct patterns (Figure 7.1): inequality rose between 1989 and
the mid-1990s and declined between the mid-1990s and 2010.1 All in all,
the Gini coefficient for per-capita (disposable monetary) income2 rose from

*
The authors are grateful to Giovanni Andrea Cornia, Richard Freeman, and Stefan Klasen,
as well as the rest of the New York (December 2010) and Buenos Aires (September 2011) work-
shop meeting participants for their very useful comments. The authors also wish to thank Alma
S. Santillan as well as Mayra Paulina Salazar and Emily Travis for their excellent research assistant-
ship. All errors and omissions remain the authors’ sole responsibility.
1
For our analyses we use information from the National Survey of Household Incomes and
Expenditures (in Spanish, Encuesta Nacional de Ingresos y Gastos de los Hogares, ENIGH) for 1989,
1992, 1994, 1996, 2000, 2006, 2008, and 2010. Although the 1989 survey is not entirely compara-
ble with the subsequent surveys, we use it to present results related to the factors behind the rise
in inequality between 1989 and 1994 (see also Annex I of Chapter 2).
2
Income includes labour income and non-labour income. The former includes all the income
that is reported as labour income in ENIGH, including labour income from self-employment.
Non-labour income includes income from own businesses, income from assets (including capital
gains), pensions (public and private), and public transfers (Oportunidades and Procampo) and pri-
vate transfers (e.g. remittances) as well as––when indicated––non-monetary income (imputed rent
on owner-occupied housing and consumption of own production, common in poor rural areas).
The surveys capture income net of taxes and contributions to social security and include govern-
ment and private transfers (remittances). Official poverty measures in Mexico use net current
income; that is, capital gains and gifts and in-kind transfers to other households are subtracted
from current total income. Current monetary income, the concept used in the decomposition of
inequality by source presented here, does not include non-monetary income and consumption of
own production (common in poor rural areas) and excludes capital gains.

140

oxfordhb-9780198701804-007.indd 140 9/27/2013 7:36:06 AM

 The Rise and Fall of Income Inequality in Mexico

0.66

0.63

0.6

0.57
Gini

0.54

0.51

0.48

0.45
1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year

Total Disp. Income Disp. Monetary Income

Labour income Hourly Wage

Figure 7.1. Evolution of the Gini coefficient, 1989–2010, Mexico

Notes:
a) Total disposable income includes labour and non-labour monetary income (net of direct taxes
and contributions to social security), transfers (private and public), and non-monetary income
(imputed rent for owner’s occupied housing, gifts in kind, and own consumption).

b) Disposable monetary income excludes non-monetary income.

c) Hourly wage is equal to monthly labour income over weekly hours of work times 4.33. Hourly
wage inequality is calculated for individuals 18–65 years old with positive income, and it includes
labour income from wages and self-employment.

d) Following standard practice, households whose head reported zero labour incomes are
excluded. Results, however, are similar if we include all households. The latter are shown in the
Statistical Appendix in Campos, Esquivel, and Lustig (2012).

e) Differences between Gini coefficients are statistically significant for the pairs: 1994–2006;
1994–2010; but not for 1989–1994 and 2006–2010.

f) There is Lorenz dominance between 1994–2006 and 1994–2010; and no Lorenz dominance
between 1989–94 and 2006–10.

g) Results are similar if we use other inequality measures such as the Theil index. See Statistical
Appendix in Campos, Esquivel, and Lustig (2012).
Source: Authors’ calculations based on ENIGH, several years.

141

oxfordhb-9780198701804-007.indd 141 9/27/2013 7:36:07 AM

 Recent Inequality Changes in Latin American Countries

0.548 to 0.571 between 1989 and 1994, and declined to 0.510 in 2010.3 The
period of declining inequality can also be divided in two: 1994–2006, when
the decline in inequality was pronounced (the Gini fell from 0.571 to 0.512);
and 2006–10, when the decline in inequality lost its momentum.4
Although during the period under analysis there were important macro-
economic fluctuations (a financial crisis in 1994/95, a steep recovery imme-
diately after that, a long but mild recession in 2000–03, and a deep GDP
contraction in 2009), most of the changes in inequality observed in Mexico
in this period can be linked to changes in the labour market and, more specifi-
cally, to changes in the wage ratio between skilled and unskilled workers. In
fact, Esquivel, Lustig, and Scott (2010) show that changes in labour income
and non-labour income inequality were equalizing for the period 1996–2006
and that the decline in labour income inequality was by far the most impor-
tant proximate determinant of the observed decline in overall inequality.5
Given the importance of labour market inequality dynamics in explaining
the trend in overall inequality, this chapter concentrates on analysing the
more ‘fundamental’ determinants of labour income inequality. In particular,
it examines the role of market forces (relative demand and supply of labour
by skill) and institutional factors (minimum wages and unionization rate) in
explaining changes in the distribution of hourly wages. It also extends the
analysis to 2010. By doing so, it examines the factors that may account for
the pause in momentum in the decline in inequality between 2006 and 2010.
More specifically, this chapter applies the ‘re-centred influence function’
(or RIF) method proposed by Firpo, Fortin, and Lemieux (2009) to decom-
pose changes in hourly wages into characteristics and returns effects.6 Results
reveal that the main driver behind the rise and decline in earnings inequal-
ity is changes in returns.7 Given the prominence of the returns effect, the

3
In this study we use the Gini coefficient as our preferred measure of inequality. This measure sat-
isfies all the desirable properties of an inequality indicator: (i) adherence to the Pigou-Dalton trans-
fer principle, (ii) symmetry, (iii) independence of scale, (iv) homogeneity, and (v) (non-additive)
decomposability. Also, the Gini is decomposable by proximate determinants as well as income
sources. We use disposable monetary income per capita unless specified otherwise. Other measures
of inequality such as the Theil index show similar trends as those described in the text. See the
Statistical Appendix in Campos, Esquivel, and Lustig (2012). As is the case with practically all ine-
quality estimates based on household surveys, the Gini coefficients presented here are probably an
underestimation of ‘true’ levels of inequality because of the significant under-reporting of incomes
and consumption at the top of the distribution (see in this respect Section 1.2.1 of Chapter 1).
4
The years 1996 and 2008 are atypical because the country was experiencing crises. In this
chapter we do not attempt to explain which factors determine inequality dynamics during a crisis.
5
The reduction in labour income inequality (leaving out the interaction terms) accounted for 87.1
per cent of the decline in inequality in 1996–2000 and for 65.5 per cent of the decline in 2000–06.
6
Although the RIF procedure was published in 2009, there have been several papers employing it
(on China, the UK, and Vietnam, for example). For references see Campos, Esquivel, and Lustig (2012).
7
In fact, changes in characteristics were un-equalizing during the period of declining inequal-
ity (1994–2006) in spite of the reduction in the Gini coefficient for education. This suggests a
persistence of what Bourguignon, Ferreira, and Lustig (2005) call the ‘paradox of progress’, which
Legovini, Bouillon, and Lustig (2005) observe in Mexico for the period 1984–94.

142

oxfordhb-9780198701804-007.indd 142 9/27/2013 7:36:08 AM

 The Rise and Fall of Income Inequality in Mexico

study proceeds in turn to analyse the determinants of the evolution of rela-

tive returns.
Changes in returns can be due to changes in the relative demand and
supply of workers of different characteristics and/or changes in institutional
factors such as the minimum wage and the unionization rate. We apply
the methodology proposed by Bound and Johnson (1992) to shed light on
which factors were predominant. The results suggest that institutional fac-
tors and the increase in relative demand for skilled workers (workers with
high school education and above) explained the increase in the inequality
of hourly wages between 1989 and 1994. This result is consistent with the
findings of a large body of existing research (see, for example, Hanson and
Harrison 1999; Bosch and Manacorda 2010). Institutional factors, however,
did not account for the decline in wage inequality between 1994 and 2006.
The evidence suggests that wage inequality fell because the supply of skilled
workers outpaced demand. The slightly rising trend in wage inequality dur-
ing 2006 and 2010 appears to be the consequence of a weakening in the
relative demand for low-skilled workers (workers with secondary education
or less).
Another factor behind the decline in overall inequality was the decline in
non-labour income inequality (Esquivel, Lustig, and Scott 2010). Non-labour
income is a very heterogeneous category. It includes all forms of income from
capital (although grossly under-reported in household surveys), pensions
from contributory systems, private transfers (remittances, in particular), and
government transfers. The importance of government transfers as an equal-
izing factor has risen considerably over time. The fiscal incidence analysis by
López-Calva et al. (2012) also underscores the growing importance of govern-
ment transfers for reducing inequality and poverty.
Figure 7.2 shows the results of applying the Lerman and Yitzhaki (1985)
decomposition.8 The contribution of income from ‘capital’ (own business,
income from property, financial income, and contributory pensions), as
expected, is always inequality-increasing whereas remittances and govern-
ment transfers are always inequality-reducing. Contribution of government
transfers is higher than that of remittances and it has grown significantly
over time. Income from capital represents, roughly, 20 per cent of total
income; income from remittances and government transfers, the remaining
20 per cent.
Labour income, which represents more than 60 per cent of total income,
does not show a definite pattern. It was inequality-increasing in 1994 and
very much so in 2010, but it was inequality-reducing in 2000, 2006, and

8
For details see Campos, Esquivel, and Lustig (2012).

143

oxfordhb-9780198701804-007.indd 143 9/27/2013 7:36:08 AM

 Recent Inequality Changes in Latin American Countries

Marginal Effect on Gini Coefficient by Income Source

0.05

0.04

0.03

0.02

0.01

0.00

–0.01

–0.02

–0.03

–0.04
Labour Income Own Transfers Remittances
–0.05
Bus.+Rents+Pensions
–0.06

1994 2000 2004 2006 2010

Figure 7.2. Decomposition of overall inequality, 1994, 2000, 2004, 2006, and
2010, Mexico
Note: Income is total current household monetary disposable (after direct taxes, contributions to
social security, and cash transfers) income in per-capita terms.
Source: Authors’ calculations based on ENIGH, several years.

2004.9 Between 1994 and 2006, the Gini coefficient of labour income fell,
while the correlation of labour and total income remained basically con-
stant. Between 2006 and 2010 there was practically no change in the Gini
coefficient of both labour and total income, but the correlation between
them increased. The latter accounts for the fact that labour income became
un-equalizing in 2010.
Given the prominent role played by labour income inequality in account-
ing for the evolution of overall inequality, below we focus on analysing the
determinants of earnings inequality. In particular, we analyse the deter-
minants of inequality in hourly wages (where ‘hourly wages’ means the
hourly remuneration of both employees and the self-employed), since
labour income inequality also reflects decisions to participate in the labour
market not examined here.

9
These results are slightly different from those presented in Esquivel (2011) and Esquivel, Lustig,
and Scott (2010), due to revisions in the data and in the definitions of income.

144

oxfordhb-9780198701804-007.indd 144 9/27/2013 7:36:08 AM

 The Rise and Fall of Income Inequality in Mexico

7.2 Determinants of Earnings Inequality: The

Contribution of Characteristics and Returns

As observed in Figure 7.1, wage inequality (measured by the Gini for hourly
wages) rose between 1989 and 1994. After 1994 there was a clear decline.
This process stops in 2006; since then, wage inequality has risen slightly in
2010.10 In this section, we analyse the main determinants of the observed trends
in wage inequality. We do this by applying the decomposition methodology
proposed by Firpo, Fortin, and Lemieux (2009).
Wage inequality is affected by two main factors: the distribution of (observ-
able and unobservable) characteristics of workers (e.g. education, experience,
gender, etc.) and the returns to those characteristics. Workers’ characteristics,
in turn, are affected by ‘fate’ (e.g. gender, race, talent, and so on), households’
decisions (e.g. to enrol in school), and policy (e.g. expanding access to edu-
cation). Returns to households’ characteristics depend on market forces (i.e.
demand and supply of workers of different skills and experience) and institu-
tional/policy factors (e.g. minimum wage policy and the unionization rate).
As one can observe in Figure 7.3, both workers’ returns and characteristics
changed between 1989 and 2010.11 The evolution of returns (panel A) follows
an inverted-U at least up until 2006. After 2006, returns to college-educated
workers begin to rise. Panel B shows that the proportion of workers with sec-
ondary, high school, and college degrees (incomplete primary and no educa-
tion) rose (declined) steadily and the relative supply of college graduates rose
faster after 1998. Measured by the Gini, inequality in the distribution of years
of schooling for Mexican workers (ages between 25 and 65) declined from
0.444 in 1989 to 0.324 in 2008 (on this, see also Chapter 15).12
We now proceed to quantify the contribution of changes in characteristics
and changes in returns to the observed changes in wage inequality. In partic-
ular, we decompose the change in log hourly wages into characteristics (also
called quantity or composition) effects and returns (also called price) effects.
Given the trends observed in Figure 7.3, we would expect the contribution of
returns to be un-equalizing between 1989 and 1994 and equalizing between

10
Results are robust for other inequality indexes. In sum, the results point out a downward trend
in labour income inequality, at least up to 2006. Since then, inequality has remained relatively
stable with a small increase in inequality by 2010, depending on how we measure inequality. We
also did the calculations using the labour force survey for the period 2005–10 (ENOE) and the
results are robust across both surveys.
11
Panel A presents the relative returns and Panel B the relative supply. Relative returns (with
respect to primary or less) are obtained from a regression of log hourly wages against dummies
of education groups (secondary, high school, and college) and control variables such as age and
geographic dummies. See Statistical Appendix in Campos, Esquivel, and Lustig (2012).
12
See Socio-Economic Database for Latin America and the Caribbean (SEDLAC), available
at: <http://sedlac.econo.unlp.edu.ar/eng/statistics-detalle.php?idE=37>.

145

oxfordhb-9780198701804-007.indd 145 9/27/2013 7:36:08 AM

 Recent Inequality Changes in Latin American Countries

Relative returns

2
Relative Returns (wrt Primary or less)

1.5

0.5

1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year

Secondary High School

College

Relative supply

0
Relative Supply (wrt Primary or less)

–0.5

–1

–1.5

–2

–2.5

1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year

Secondary High School

College

Figure 7.3. Relative returns and relative supply, 1989–2010, Mexico

Notes: Sample restricted to workers 18–65 years old. Panel A plots relative returns of education
groups with respect to primary or less. Panel B plots relative supply (in logs) of education groups
with respect to primary or less. Primary or less refers to individuals with less than secondary
(9 years of schooling), secondary refers to individuals with equal to or more than 9 and less than
12 years of schooling, high school refers to individuals with equal to or more than 12 and less
than 16 years of schooling, and college refers to individuals with at least 16 years of schooling.
Source: Calculations by the authors using ENIGH.

146

oxfordhb-9780198701804-007.indd 146 9/27/2013 7:36:08 AM

 The Rise and Fall of Income Inequality in Mexico

1994 and 2006. In contrast, the effect of changes in the composition of char-
acteristics cannot be inferred ex ante.
Although there was significant educational upgrading and the distribution
of the stock of education became more equal over the entire period under
study, whether this change was equalizing or un-equalizing depends on the
extent of convexity in the returns to education and at what point of the edu-
cation equalization process the country found itself. Bourguignon, Ferreira,
and Lustig (2005) were among the first to notice that a reduction in the ine-
quality of education––in the presence of increasing returns to education––
could lead to a rise in earnings inequality. They call this result the ‘paradox
of progress’ alluding to the fact that a more equal stock of education can
be inequality-increasing (at least during part of the educational upgrading
process) if the returns to education increase at an increasing rate with the
level of attainment (convexity in the returns). As noted in Chapter 15 and by
Gasparini et al. (2011), the ‘paradox of progress’ has been quite a pervasive
phenomenon in Latin American labour markets in the last couple of decades.

7.2.1 Decomposing Wage Inequality into Characteristics and Returns

Effects: An Application of the Re-centred Influence Function (RIF)
Procedure (1989–2010)
There are many decomposition procedures that are employed in the litera-
ture (see the excellent review by Firpo, Fortin, and Lemieux 2011). Most of
them rely on a Oaxaca-Blinder (OB) type of decomposition.13 In this chapter,
we employ the ‘re-centred influence function’ (RIF) procedure proposed by
Firpo, Fortin, and Lemieux (2009) to decompose effects into characteristics or
composition and returns effects.
The RIF procedure is very similar to the typical OB decomposition.14 The
main difference is that the dependent variable, Y, is replaced by the ‘re-centred
influence function’ (RIF).15 Firpo, Fortin and Lemieux (2009) demonstrate
that the RIF procedure is equivalent to a simple unconditional quantile

13
We can divide the decomposition into four groups: (i) reweighting procedures (DiNardo,
Fortin and Lemieux 1996), (ii) residual-imputation procedures (Almeida dos Reis and Paes de
Barros 1991; Juhn, Murphy and Pierce 1993), (iii) quantile decomposition procedures (Machado
and Mata 2005), and (iv) re-centred influence function (RIF) procedures (Firpo, Fortin and Lemieux
2009; see also Chapter 12 for a discussion of the decomposition methods).
14
See the papers by Firpo, Fortin, and Lemieux (2009, 2011) for more details of the RIF procedure.
15
Define RIF(v,y) as the re-centred influence function with distributional statistic of interest v(Fy)
and observed wage y. Then it can be shown that RIF(v,y)= v(Fy)+IF(v,y), where IF denotes the influ-
ence function such that ∫ RIFI v ( Fy ) . For the case of quantiles, it can be shown that the influence
τ− { τ } . Each statistic v(F ) refers to a specific quantile in the
function is equal to ( τ ,Y ) = y
fY ( τ )
distribution of Y or to the Gini coefficient or the variance.

147

oxfordhb-9780198701804-007.indd 147 9/27/2013 7:36:08 AM

 Recent Inequality Changes in Latin American Countries

regression. They show that E ⎡⎣ RIF (v y )|X ⎤⎦ Xβv , where the coefficient βv
represents the marginal effect of X on the dependent variable statistic v.16
Once we estimate the parameter β for each year in our sample, we
v

apply a Oaxaca-Blinder decomposition. In other words, we estimate

(
v (Yt ) v (Ys ) = β s X t
v
)⎛ v v⎞
X s + X t β t − β s where t is the final year and s is the
⎝ ⎠
initial year. In our application, we set up the initial years as 1989, 1994, and
2006 and the final years as 1994, 2006, and 2010, respectively. As is typical
v
( )
in an OB decomposition, the term β s X t X s refers to the characteristics

effects and the term X t ⎛ β t β s ⎞ refers to the return or price effects to observ-
v v
⎝ ⎠
able characteristics included in X and also, unobservable ones (which is why
this term is often referred to as the ‘unexplained component’). We use as
reference the wage distribution in the initial year (for each decomposition).
Figure 7.4 shows the decomposition for quantiles 1, 2, . . ., 99. In other
words, we estimate the RIF procedure in every quantile and obtain the dif-
ference in the average wage for each quantile and then the part attributed to
characteristics and to returns. The figure includes three panels for different
periods. Panel A (1989–94) shows that inequality increased during the period.
In this period, observable characteristics explained little of the increase in
inequality, given that the part explained by characteristics is a flat line. The
increase in inequality was mostly due to returns as shown by the upward
sloping shape of the ‘effects of returns’ curve.
Panel B (1994–2006), on the other hand, shows that inequality decreased
during the period. Wages for low-earning individuals rose while those for
richer individuals declined. Interestingly, the effects of characteristics (edu-
cation, experience, female, and urban) were inequality-increasing. In other
words, if returns to characteristics had been equal to their 1994 level, the
change in characteristics in the population (in spite of the equalization of
education) would have increased inequality. This points to a persistence of the
‘paradox of progress’ found for Mexico (1984–94) by Legovini, Bouillon, and
Lustig (2005). Hence, the driving force behind the decline in wage inequality
between 1994 and 2006 must have been the effects of returns. As shown in

ν=0 5
16
For example, if v represents quantile 0.50, then β represents the effect of X on the wage
quantile 0.50. It can also be applied to scalar indicators of inequality such as the Gini or the vari-
m
ance. In order to estimate the RIF regression, we first estimate the sample RIF
I ( ,Y ) . In practice,
we follow the ado file rifreg in Stata published by Firpo, Fortin, and Lemieux (2011) provided
by N. Fortin <http://faculty.arts.ubc.ca/nfortin/datahead.html>. The RIF dependent variable is
estimated using kernel methods. We use the following explanatory variables: dummy variables of
female, urban, education categories and a cubic polynomial in age. We also estimated a more flex-
ible model that included interactions among all variables; however, the difference in explained
and unexplained components was minimal.

148

oxfordhb-9780198701804-007.indd 148 9/27/2013 7:36:09 AM

 The Rise and Fall of Income Inequality in Mexico

(a) 1989–94

0.7

Log wage effects

0.4

0.1

–0.2

–0.5
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95100
Quantile

Total differential Effects of characteristics

Effects of returns

(b) 1994–2006

0.7
Log wage effects

0.4

0.1

–0.2

–0.5

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95100
Quantile

Total differential Effects of characteristics

Effects of returns

(c) 2006–10

0.7
Log wage effects

0.4

0.1

–0.2

–0.5
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95100
Quantile

Total differential Effects of characteristics

Effects of returns

Figure 7.4. Decomposition of differences in the distribution of earnings, 1989–2010,

Mexico
Notes: Total differential is the total change in hourly wages (in logs); Effects of characteristics and
Effects of returns are the portions that one can ascribe to changes in characteristics and returns
respectively.
Source: Calculations by the authors using ENIGH.

149

oxfordhb-9780198701804-007.indd 149 9/27/2013 7:36:12 AM

 Recent Inequality Changes in Latin American Countries

Panel B, the effects of returns contributed to equalize the earnings distribu-

tion by such an amount that they compensated for the inequality-increasing
effects associated with the changes in characteristics. Although we do not dis-
aggregate the returns into various components, this result is consistent with
the fall in the relative returns to education shown in Figure 7.3.
Panel C (2006–10) shows that although changes in hourly wages were prac-
tically nil across most of the distribution, individuals at the bottom suffered
declines in wages. Observable characteristics do not contribute to an expla-
nation for the changes in inequality in this period. However, and in contrast
with the 1994–2006 period, the decline in relative returns to low-wage work-
ers accounted for their decline in relative wages.
In sum, these results suggest that the driving force behind the rise (1989–
94), decline (1994–2006), and slight increase (2006–10) in wage inequality
was changes in relative returns. Our next task is to determine which factors
explain the behaviour of relative returns. We shall concentrate on the rela-
tive returns to skill because they experienced prominent changes, as shown
in Figure 7.3, Panel A.

7.3 Determinants of Relative Returns: The Role of Demand,

Supply, and Institutional Factors

The wage structure is affected by demand and supply of workers of different

skills and by institutional factors such as the minimum wage and unions.
Labour demand by skill, in turn, is primarily affected by the characteristics of
technical change and international trade. The composition of labour supply
is determined, to a large extent, by the characteristics of educational upgrad-
ing. Figure 7.5 plots the relative returns and relative supply of workers with
high school education or more against workers with secondary or less. The
left y-axis shows the relative returns and the right y-axis the relative supply
in logs. The increase in relative supply is larger for the period 1996/98–2010
than for the period 1989–1996/98. The increase in relative supply for the
period 1989–98 is approximately 20 per cent, while for the period 1998–2010
it is approximately 54 per cent. Inequality measured as the relative returns for
workers with at least high school education, on the other hand, increases for
the period 1989–94 and it clearly declines for the period 1998–2010.
Following Bound and Johnson (1992), if increases in supply are larger than
increases in demand—everything else being equal––then we expect relative
returns to fall. For the period 1989–94 we observe both an increase in rela-
tive supply and a rise in relative returns for workers with tertiary education.
Hence, either demand outpaced supply for skilled labour, or institutional fac-
tors disfavoured the unskilled, or both. The rapid increase in wage inequality

150

oxfordhb-9780198701804-007.indd 150 9/27/2013 7:36:12 AM

 The Rise and Fall of Income Inequality in Mexico

1.1 –0.8

–1
1
Relative Returns

Relative Supply
–1.2

0.9

–1.4

0.8
–1.6
1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year

Relative Retuns Relative Supply

Figure 7.5. Relative returns and relative supply, 1989–2010, Mexico (high school and
more vs. secondary or less)
Notes: Sample restricted to workers 18–65 yrs old. Relative returns are obtained from a regression
of log hourly wages. Relative supply is equal to the log of the ratio of proportion of workers with
high-school or college education over the proportion of workers with secondary or less. For more
details, see Campos, Esquivel, and Lustig (2012).
Source: Calculations by the authors using ENIGH.

that occurred in Mexico between the mid-1980s and the mid-1990s has been
the subject of a fairly large body of research.17 The main conclusions are that
institutional factors as well as skill-biased demand explain the observed trend.
Further details are discussed in the last section of the chapter.
What about the period 1994–2006 when wage inequality declined? In
Figure 7.5 we observe that the relative supply of skilled workers rose while the
relative returns declined. This means that either supply outpaced demand,
institutional factors moved in favour of the unskilled, or both. Figure 7.6
shows the evolution of the real minimum wage and the unionization rate
for the period 1988–2010. Panel A includes the monthly index of the real
minimum wage using December 2010 as the base period. The real minimum
wage fell by 50 per cent between 1988 and 1996. However, after 1996 the
real minimum wage was fairly stable. Hence, it is unlikely that the minimum

17
There are plenty of references that analyse the determinants of changes in inequality. A sum-
mary of the literature can be found in Bouillon, Legovini, and Lustig (2003), Campos (2010),
Campos, Esquivel, and Lustig (2012), Esquivel (2011), Esquivel and Rodríguez-López (2003), and
Legovini, Bouillon, and Lustig (2005).

151

oxfordhb-9780198701804-007.indd 151 9/27/2013 7:36:12 AM

 Recent Inequality Changes in Latin American Countries

(a) Real minimum wage index

(December 2010=100)

200
Real Minimum Wage (December 2010=100)

180

160

140

120

100

1988m1 1992m1 1996m1 2000m1 2004m1 2008m1 2010m12

Year

(b) Unionization rate

0.2

0.18
Unionization Rate

0.16

0.14

0.12

0.1
1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year

ENIGH ENOE

Figure 7.6. Real minimum wage and unionization, 1988–2010, Mexico

Notes: Real minimum wage index is obtained from Comisión Nacional de Salarios Mínimos
<http://www.conasami.gob.mx/> and the unionization rate is obtained from two different sur-
veys. ENIGH provides union information up to 2006. ENOE (Encuesta Nacional de Ocupación y
Empleo) provides union information for the period 2005–10.

wage affected the wage structure for the period after 1994. While there is a
marked decline in unionization between 1989 and 1996, there was no major
change after 1996, although there appears to be a slight decline in unioniza-
tion after 2005 (approximately 1 percentage point). The minimum wage may
affect the distribution of wages if the minimum wage is binding, because this

152

oxfordhb-9780198701804-007.indd 152 9/27/2013 7:36:13 AM

 The Rise and Fall of Income Inequality in Mexico

could result in stable real wages at the bottom even if wages higher up in the
distribution experience a decline. Existing evidence suggests that the mini-
mum wage is currently not and has not been binding since the mid-1990s.
Following Bosch and Manacorda (2010), Figure 7.7 shows the wage distribu-
tions in 1989 and 2010 for the urban sector once we subtract the median
wage. The vertical line is the value of the minimum wage minus the median
wage. The figure shows that the minimum wage could have been (slightly)
binding in 1989 but not in 2010.
In sum, it appears that institutional factors such as the minimum wage and
the unionization rate did not play a role in explaining the trends in relative
wages/returns during 1996–2010. The evolution of relative wages/returns in
this period seems to be associated with how the demand and supply of labour
of different skills changed over time. For the period 1994–2006, the fall in
relative returns appears to have occurred because the supply of high-skilled
workers outpaced demand. Since the supply of skilled workers continued to
increase during 2006–10, the rise in relative returns suggests that either rela-
tive demand for skilled labour outpaced supply or, that the relative supply

0.75
Density

0.5

0.25

0
–3 –2 –1 0 1 2 3
Log wage - Median

1989 2010

Figure 7.7. Wage distribution with respect to median wage, 1989 and 2010, Mexico
Notes: Calculations by the authors using labour force surveys (ENEU and ENOE) for the urban sec-
tor and for full-time workers (more than 25 hours per week). Wage distributions using monthly
earnings. Vertical lines show the log of the minimum wage assuming full-time work during the
month minus the median monthly wage.
Source: Calculations by the authors using labour force surveys (ENEU and ENOE).

153

oxfordhb-9780198701804-007.indd 153 9/27/2013 7:36:13 AM

 Recent Inequality Changes in Latin American Countries

of unskilled workers outpaced demand.18 We now attempt a more rigorous

estimation and account of demand and supply factors

7.3.1 The Effect of Demand and Supply on Relative Wages: An

Application of the Bound and Johnson Method (1989–2010)
In order to examine the effect of supply and demand on relative wages, we
follow the Bound and Johnson (1992) method.19 Based on the evidence
presented in Figures 7.5 and 7.6, and as discussed above, we assume that
non-competitive factors (i.e. minimum wages and unionization rate) are not
important during the 1994–2010 period and ascribe the observed trends in
relative wages by skill to demand and supply factors alone.
Assuming a simple CES production function with elasticity of substitution,
σ, constant across skills, it is possible to determine the effect of supply and
demand on relative wages.20 In particular, it is possible to show that the rela-
tive wage of workers with at least a high-school diploma (wC) in terms of the
wage of workers with at most secondary education (wS) can be expressed in
terms of its increase in demand and supply:

⎛ −C ⎞
W 1 1
Δ% ⎜ − S ⎟ = Δ%
Δ %( d) − l )+ ξ
Δ%(Supply
⎜ ⎟ σ σ
⎝W⎠

The residual term ξ contains the effect of skill-biased technical change and
other non-competitive factors. As the unionization rate and the real minimum
wage were fairly constant during 1994–2006, we assume non-competitive
factors are negligible. In order to make the simulation simpler, we simulate
changes only in supply and assign the full residual to demand and skill-biased
technical change (which affects demand, of course). The supply component
is equal to the relative increase of workers with at least high-school educa-
tion divided by workers with at most secondary education. Table 7.1 shows

18
Using ENOE for the period 2006–10, we find that the relative returns of college-educated
workers against workers with primary or less declined 0.01 points. However the decline in returns
was larger for high-school-educated workers and workers with secondary education. Hence, the
result of the slowdown in returns for college-educated workers is robust to the selection of the
microdata: ENIGH and ENOE.
19
We attempt to estimate a model similar to Bound and Johnson (1992) and Manacorda,
Sánchez-Páramo, and Schady (2010). However, as pointed out by Manacorda, Sánchez-Páramo,
and Schady (2010), the relevant elasticities of substitution for the case of Mexico cannot be pre-
cisely estimated. In order to estimate the structural parameter σ, these authors use a sample of
workers from Argentina, Brazil, Chile, Colombia, and Mexico; they mention that ‘Mexico does
not really contribute to the identification of the regression parameters’ (page 314, footnote 1).
20
See formula (3) in page 377 and formula (A8) in page 390 of Bound and Johnson (1992).

154

oxfordhb-9780198701804-007.indd 154 9/27/2013 7:36:13 AM

 The Rise and Fall of Income Inequality in Mexico

Table 7.1. Effects of relative labour supply on relative wage, 1989–2010, Mexico

Change Returns Supply = Rest

Panel A. σ=1
1989–94 0.240 0.111 0.351
1994–2006 –0.310 0.474 0.164
2006–10 0.020 0.154 0.174
Panel B. σ=2
1989–94 0.240 0.055 0.295
1994–2006 –0.310 0.237 –0.073
2006–10 0.020 0.077 0.097

Source: Authors’ estimates based on ENIGH, several years.

the results of the simulation assuming an elasticity of substitution of 1 and 2

which is the consensus in the literature (Bound and Johnson 1992; Katz and
Autor 1999).
Consistent with previous research findings, Table 7.1 suggests that
changes in relative supply had a small effect on relative wages in the period
between 1989 and 1994. Most of the changes for that period, then, have to
be explained by changes in demand and institutional factors, as discussed
above. The relative contribution of market versus institutional factors,
however, cannot be clearly disentangled.
For the period of declining relative wages (1994–2006), Table 7.1 shows
that relative returns declined by 31 per cent for the period 1994–2006.
However, had nothing changed in the same period other than relative
supply, then relative returns would have declined by as much as 47 per
cent. A key issue arises, however, depending on the value we assume for
the elasticity of substitution. In Panel A (σ = 1), relative demand shows a
steady growth for the period 1994–2010. In Panel B (σ = 2), relative demand
declined for skilled workers (high-school and college) during the period
1994–2006 and started to rise again for the period 2006–10. Taking the
median value of the elasticity of substitution (not shown on Table 7.1), the
patterns show a slowdown in demand for the period 1994–2006, and then a
rise for the period 2006–10. Hence, the rapid increase in relative supply was
a key component in explaining the reduction in relative wages, but only up
to 2006. In recent years, demand patterns appear once again to benefit the
highly skilled to a larger degree. Based on the analysis presented in Section
7.3 and Figure 7.4 (Panel C), it would appear that during the 2008/09 reces-
sion and its aftermath, relative demand for low-wage/low-skilled workers
declined the most.

155

oxfordhb-9780198701804-007.indd 155 9/27/2013 7:36:14 AM

 Recent Inequality Changes in Latin American Countries

7.4 Cash Transfers and Inequality

In Table 7.2, one can observe the changes in total disposable income per
capita21 as a result of government transfers. The calculations presented in
this table are the result of a standard incidence analysis of government trans-
fers.22 As one can see, the contribution of government cash transfers to the
reduction in inequality and poverty was almost nil in 1996, rose in 2000,
and became more significant, especially for poverty reduction, in 2010.

Table 7.2. The impact of cash transfers on inequality and poverty, 1996, 2000, and
2010, Mexico

Net market income Disposable income

1996 Gini 0.522 0.520

% change with respect to net market –– –0.4%
income
Headcount index ($2.5 PPP) 30.2% 29.9%
% change with respect to net market –– –1.0%
income
2000 Gini 0.544 0.539
% change with respect to net market –– –0.9%
income
Headcount index ($2.5 PPP) 22.1% 21.6%
% change with respect to net market –– –2.3%
income
2010 Gini 0.503 0.495
% change with respect to net market –– –1.7%
income
Headcount index ($2.5 PPP) 13.8% 11%
% change with respect to net market –– –20.1%
income

Notes: Income variables here include monetary and non-monetary components which explain the bulk of the
difference between the Gini coefficients reported here and in the first paragraph of the chapter and the Statistical
Appendix in Campos, Esquivel, and Lustig (2012). The remaining differences are due to rounding errors.

Net market income is total market income minus direct taxes and contributions to social security.

Disposable income is net market income plus government transfers (private transfers and contributory pensions are
included in market income).

Source: López-Calva, Lustig, and Scott (2012).

21
The differences between the Ginis here and those presented in the first paragraphs of this
study are due to the fact that there we include information on monetary income only while we use
total income here. Total income includes monetary income plus auto-consumption and imputed
rent for owner’s occupied housing.
22
For details, see Lustig et al. (2011) and López-Calva et al. (2012). Unfortunately, due to limita-
tions of the data, it was not feasible to conduct this analysis for years prior to 1996. However, 1996
is the year before the cash transfer programme Progresa was launched. Hence, the results for 1996
can be used as a baseline.

156

oxfordhb-9780198701804-007.indd 156 9/27/2013 7:36:14 AM

 The Rise and Fall of Income Inequality in Mexico

Most of this change is due to Progresa, the flagship conditional cash trans-
fer programme launched in 1997 (which changed its name to Oportunidades
in 2002).
Oportunidades is a federal conditional cash transfer programme that targets
rural and urban households in Mexico that fall within the extreme poverty
category. It complements traditional supply-side spending on social services
with demand-side subsidies. The programme has three components: educa-
tion, nutrition, and health. The education component grants cash transfers
based on school attendance, high school completion, and the need for school
supplies. The nutrition and health components offer cash and in-kind trans-
fers (nutritional supplements, vaccinations, preventative treatments, and so
forth), based on regular visits to a health clinic. The average monthly transfer
is about US$35 and estimated total transfers are equivalent to, on average,
25 per cent of eligible rural households’ average monthly income. The pro-
gramme’s size is significant in terms of the number of beneficiaries, yet it
is inexpensive in terms of cost. By the end of 2010, Progresa/Oportunidades
granted benefits to 5.8 million families (about 27 per cent of the Mexican
population). Its budget in 2010 equalled 0.48 per cent of GDP (compared with
0.02 per cent in 1997), and it commanded close to 2.5 per cent of the pro-
grammable public expenditure budget. Impact evaluation studies have found
that the programme has had positive impacts on education and health.23
All in all, Progresa/Oportunidades transformed the broadly neutral distri-
bution of government spending on food subsidies into a highly progressive
one: the share benefiting the poorest decile increased from 8 to 33 per cent
between 1994 and 2000.24 Beyond its effects on education, health, and nutri-
tion, Progresa/Oportunidades has had a positive impact on poor households’
consumption, thereby helping to reduce poverty and inequality in Mexico.25

7.5 Concluding Remarks: The Rise and Fall of Income

Inequality and Policy Regimes

Previously we identified three episodes in inequality dynamics in Mexico: a

period of rising inequality (1989–94); a period of declining inequality (1994–
2006); and a period in which the decline in inequality lost its momentum
(2006–10). These periods coincide roughly with two broad policy regimes
(Table 7.3). Between 1989 and 1994, the policy regime was characterized by
intense and widespread market-oriented reforms (with trade liberalization

23
For citations see Campos, Esquivel, and Lustig (2012) and Lustig (2011).
24
See Scott (2009).
25
For references, see Campos, Esquivel, and Lustig (2012).

157

oxfordhb-9780198701804-007.indd 157 9/27/2013 7:36:14 AM

 Recent Inequality Changes in Latin American Countries

and privatizations taking the lead), dismantling of price supports and gen-
eralized subsidies, and reductions in the minimum wage and unionization
rates (similar effects were observed in the other country case studies in this
volume). After 1994, the policy regime was characterized by a paucity of
structural reforms, strategic integration with the rest of the world (of which
the salient example is the North American Free Trade Agreement or NAFTA),
and the introduction of large-scale (in terms of beneficiaries) cash transfer
programmes. Minimum wages became non-binding and the unionization
rate remained low. What, if any, might be the connection between the policy
regimes and inequality outcomes?
Our analysis indicates that the rise in overall inequality between 1989 and
1994 is accounted for, to a large extent, by the rise in labour income inequal-
ity. This, in turn, is associated with the increase in relative returns for skilled
workers (those who hold a high-school diploma or more). The increase in
the skilled–unskilled wage gap coincided with the unilateral trade liberaliza-
tion that started in the mid-1980s (Table 7.3; see also Chapter 11). In that
sense, the evolution of Mexico’s wage inequality was unexpected; Mexico
had an abundance of relatively unskilled labour (at least from the perspective
of its main trade partner, the United States), and standard theories of trade
predicted exactly the opposite pattern (that is, a reduction in the skilled–
unskilled wage ratio).26
Why did trends in relative wages during 1989 and 1994 contradict expec-
tations stemming from standard trade theory? First, this period also coin-
cided with labour market policies/institutional changes that disfavoured the
low-skilled: a reduction in real minimum wages and in the unionization rate
(Table 7.3 and Figure 7.6). Bosch and Manacorda (2010) find evidence that
these institutional factors were quite decisive in causing wage inequality to
rise. In addition, there is evidence that the direct and indirect impact of the
opening up of the economy (trade liberalization and foreign direct invest-
ment liberalization) contributed to the rise in the wage gap by skill. The direct
effect occurred because—contrary to expectations––some labour-intensive
sectors (such as textiles and garments) were relatively more protected under
import-substitution industrialization and were hurt by trade liberalization.27
The indirect effect manifested itself through skill-biased technical change
(though, admittedly, it is hard to disentangle which part of the latter is
induced by openness or occurs independently).
Is there a connection between the policies pursued after 1994 and the
decline in overall inequality? Again, the results of the decomposition exercise
presented in Section 7.2 and in Esquivel, Lustig, and Scott suggest that one

26
For a discussion of trade liberalization and its implications, see, for example, Lustig (1998).
27
See, for example, Hanson and Harrison (1999).

158

oxfordhb-9780198701804-007.indd 158 9/27/2013 7:36:14 AM

 The Rise and Fall of Income Inequality in Mexico

Table 7.3. Policy regimes, 1989–2010, Mexico

1989–94 1994–2010

Macro – Aftermath of 1980s debt crisis – 1995 peso crisis and recovery
– Contractionary fiscal and monetary – Fiscal discipline (balanced budget
policies–Quasi-fixed exchange law passed in 2006)
– Very low growth – Inflation-targeting by central bank
– Inflation under control starting in since 1999
1989 – Flexible exchange rate regime
– Low growth (GDP/capita growth
of around 1% annually) with some
inflation in the second half of
1990s; low inflation since around
2000
– Output contracted sharply in
2008/09 due to great recession
in US
Labour – Minimum wages and unionization – Minimum wages stable and not
rates declined markedly binding. Unionization rates stable
with a slight decline since 2005
Openness – Unilateral trade liberalization since – NAFTA comes into effect in 1994.
1985. Mexico joins GATT in 1986. Other free trade agreements
– Foreign direct investment liberalized
Other – Large scale privatizations (banks and – Social security reforms
market-oriented telecommunications)
reforms – Deregulation
– Dismantling of price support (and
other) schemes in agriculture and
elimination of general production
and consumption subsidies
Social policy – Very small-scale targeted subsidies – Targeted cash transfer
to tortilla programmes: Procampo in 1995
– Flagship anti-poverty program and Progresa in 1997. Progressa
Programa Nacional de Solidaridad changes name to Oportunidades
focused on expanding rural in 2002 and is expanded to urban
infrastructures (no targeted cash areas and includes children in high
transfer) school
– Non-contributory pensions in rural
areas in 2007 (Seventy or more)
Inequality – Increased – Declined especially between 1998
and 2004; between 2006 and
2010, decline loses momentum
and wage inequality slightly rises

Notes:

a) Progresa/Oportunidades: Launched in 1997; provides direct monetary and in-kind transfers conditional on school
attendance and health-centre visits. By the end of 2010, Progresa/Oportunidades granted benefits to 5.8 million
families (about 27 per cent of the Mexican population).

b) Procampo: Direct monetary transfer per hectare, originally set at close to US$100 per hectare to all beneficiaries
identified in the original 1993 survey on the basis of cultivation of nine basic crops. Conditional on cultivation of the
land, but after 1995 not conditional on particular crops. Administrative data: 2.39 million beneficiaries in 2008.

c) Seventy or more: Non-contributory pension. All the population of 70 years and older living in localities of
30,000 or less are eligible for this universal rural non-contributory basic pension of 500 pesos (US$37) per month.
Administrative data: 1.031 million beneficiaries in 2008.

Source: Authors’ compilation based on Lustig (2010).

159

oxfordhb-9780198701804-007.indd 159 9/27/2013 7:36:14 AM

 Recent Inequality Changes in Latin American Countries

of the most important inequality-reducing forces between 1994 and 2006

has been the evolution of labour income inequality. Note that labour income
is basically the result of multiplying hours worked by hourly wages (here
defined as including remunerations to the self-employed). It turns out that
hours worked did not change much from 1994 to 2006,28 so the change in
labour income inequality must have been caused by changes in hourly wage
inequality. Some authors have linked the reduction in wage inequality to
NAFTA. Robertson (2007), for example, suggests that Mexico’s manufacturing
workers are now complements, rather than substitutes, to US workers. He also
posits that there has been an important expansion of assembly-line activi-
ties in Mexico (maquiladoras), which has increased demand for less-skilled
workers. Campos (2010) emphasizes the supply-side explanations based on
changes in the composition of the labour force.
Between 1989 and 1994, most of the changes in the wage distribution
occurred in the upper tail of the distribution (workers with high wages and
high levels of education and experience). As was seen in Figure 7.4, Panel
A, the increase in wage inequality in those years was not caused by a (rel-
ative) decline in the wages of the low-skilled or low-experienced workers;
rather it was the result of a higher rise in the wages of the high-skilled or
high-experienced workers. In contrast, between 1994 and 2006 the reduc-
tion in wage inequality was caused by the changes in the lower tail of the
income distribution. Average wages for workers with lower levels of educa-
tion and/or fewer years of experience increased (Figure 7.4, Panel B), even
though average real and legislated minimum wages were practically flat over
this period (Figure 7.6). Average wages for higher-paid workers (high-skilled
and/or high-experienced workers), in contrast, declined between 1994 and
2006 (Figure 7.4, Panel B).
For the post-NAFTA period (after 1994), then, there are at least two (not
mutually exclusive) possible explanations: an increase in the relative sup-
ply of skilled workers and an increase in the demand for low-skilled labour
resulting from an expansion in assembly-line activities (maquiladoras) in
Mexico’s manufacturing sector. Based on our analysis presented in Section
7.4 (Table 7.1), the reduction in relative returns of the high-skilled workers
seems to be driven, primarily, by the rise in their relative supply.
The increase in the relative supply of workers with high levels of skills
reflects the significant educational upgrading of the labour force that occurred

28
Actually, between 1994 and 2006, weekly hours in all jobs fell slightly and the decline was
concentrated in low-education (poorer) workers, which would be an inequality-increasing change.
This means that the inequality-reducing changes in the distribution of hourly earnings must have
been large enough to compensate for the inequality-increasing effect of the changes in the distri-
bution of hours worked. Data on weekly hours and hourly wages are available at: <http://www.
depeco.econo.unlp.edu.ar/sedlac/>.

160

oxfordhb-9780198701804-007.indd 160 9/27/2013 7:36:14 AM

 The Rise and Fall of Income Inequality in Mexico

during this period (Figure 7.5). Part of this upgrading should be the conse-
quence of the expansionary policies in terms of access to education (on this,
see Chapter 15). However, part might also be a consequence of more indi-
viduals deciding to invest in a tertiary degree in response to the rising returns
to skill experienced between 1989 and 1994 (and, actually since 1984). This
would suggest that Mexico experienced a Tinbergean process in the sense
that skill-biased demand (due to trade liberalization and technical change)
contributed along with institutional factors to a significant increase in the
skill premium. This, in turn, could have induced individuals to invest more in
their own education by completing high-school and tertiary qualifications.
The subsequent increase in the relative supply of more educated workers
caused the skill premium and wage inequality to decline.
In sum, the results reveal the following. Relative supply only marginally
affected the wage structure during the period 1989–94. Therefore, relative
demand and institutional factors are responsible for the increase in inequal-
ity. On the other hand, after 1994 institutional factors have remained largely
unchanged. At the same time, relative supply of skilled labour (completed
high school or more) increased by more than 50 per cent and relative demand
slowed down, which resulted in lower inequality. The period 2006–10 has
seen a small increase in inequality. This is mainly due to a decrease in wages
at the bottom and not to an increase of wages at the top. Does this point to
a reversal in the wage inequality dynamics in Mexico? At this point, it is too
soon to be able to disentangle the permanent versus the temporal effects of
the recent macroeconomic crisis caused by the recession in the United States.
Finally, overall inequality has declined because non-labour income ine-
quality declined too. Our analysis and that presented in Esquivel, Lustig, and
Scott (2010) suggest that a change in social policy from general subsidies to
cash transfers targeted to the poor contributed to the decline in inequality—
especially since 2000, when the number of beneficiaries was increased.

References
Almeida dos Reis, J. and R. Paes de Barros (1991). ‘Wage Inequality and the
Distribution of Education: A Study of the Evolution of Regional Differences in
Inequality in Metropolitan Brazil’. Journal of Development Economics, 36(1): 117–43.
Bosch, M. and M. Manacorda (2010). ‘Minimum Wages and Earnings Inequality in
Urban Mexico’. American Economic Journal: Applied Microeconomics, 2(4): 12–49.
Bouillon, C., A. Legovini, and N. Lustig (2003). ‘Rising Inequality in
Mexico: Household Characteristics and Regional Effects’. Journal of Development
Studies, 39(4): 112–33.
Bound, J. and G. Johnson (1992). ‘Changes in the Structure of Wages in the
1980s: An Evaluation of Alternative Explanations’. The American Economic Review,
82(3): 371–92.

161

oxfordhb-9780198701804-007.indd 161 9/27/2013 7:36:14 AM

 Recent Inequality Changes in Latin American Countries

Bourguignon, F., F. Ferreira, and N. Lustig (2005). The Microeconomics of Income

Distribution Dynamics in East Asia and Latin America. Washington, DC: Oxford
University Press, Washington.
Campos, R. (2010). ‘Why did Wage Inequality Decrease in Mexico after NAFTA?’.
Centro de Estudios Economicos Working Paper 2010-15, El Colegio de Mexico.
Campos, R., G. Esquivel, and N. Lustig (2012). ‘The Rise and Fall of Income
Inequality in Mexico, 1989–2010’. WIDER Working Paper 2012/10.
Helsinki: UNU-WIDER.
DiNardo, J., N. M. Fortin, and T. Lemieux (1996). ‘Labour Market Institutions and
the Distribution of Wages, 1973–1992: A Semiparametric Approach’. Econometrica,
64(5): 1001–44.
Esquivel, G. (2011). ‘The Dynamics of Income Inequality in Mexico since NAFTA’.
Economia, 12(1): 155–79.
Esquivel, G., N. Lustig, and J. Scott (2010). ‘Mexico: A Decade of Falling
Inequality: Market Forces or State Action?’ In L. F. López and N. Lustig (eds),
Declining Inequality in Latin America: A Decade of Progress? Washington,
DC: Brookings Institution Press and New York: United Nations Development
Programme.
Esquivel, G. and J. A. Rodríguez-López (2003). ‘Technology, Trade and Wage
Inequality’. Journal of Development Economics, 72(2): 543–65.
Firpo, S., N. Fortin, and T. Lemieux (2009). ‘Unconditional Quantile Regressions’.
Econometrica, 77(3): 953–73.
Firpo, S., N. Fortin, and T. Lemieux (2011). ‘Decomposition Methods in
Economics’. In D. Card and O. Ashenfelter (eds), Handbook of Labour Economics.
Amsterdam: North Holland, 2–104.
Gasparini, L., S. Galiani, G. Cruces, and P. Acosta (2011). ‘Educational Upgrading and
Returns to Skills in Latin America. Evidence from a Supply-Demand Framework.’
WB Policy Research Working Paper 5921. Washington, DC: World Bank.
Hanson, G. and H. Harrison (1999). ‘Trade Liberalization and Wage Inequality’.
Industrial and Labour Relations Review, 52(2): 271–88.
Juhn, C., K. Murphy, and B. Pierce (1993). ‘Wage Inequality and the Rise in Returns
to Skill’. Journal of Political Economy, 3(3): 410–48.
Katz, L. F. and D. H. Autor (1999). ‘Changes in the Wage Structure and Earnings
Inequality’. In O. Ashenfelter and D. Card (eds), Handbook of Labour Economics, vol.
3. Amsterdam: North Holland.
Legovini, A., C. Bouillon, and N. Lustig (2005). ‘Can Education Explain Changes in
Income Inequality in Mexico’. In F. Bourguignon, F. H. G. Fereira, and N. Lustig
(eds), The Microeconomics of Income Distribution Dynamics. Washington, DC: Oxford
University Press for the World Bank, 275–312.
Lerman, R. I. and S. Yitzhaki (1985). ‘Income Inequality Effects by Income
Source: A New Approach and Applications to the United States’. Review of
Economics and Statistics, 67: 151–6.
López-Calva, L. F., N. Lustig, J. Scott, and A. Castañeda (2012). ‘Government Transfers
and Redistribution in Mexico: 1992–2010’. Work in progress. Washington,
DC: World Bank.

162

oxfordhb-9780198701804-007.indd 162 9/27/2013 7:36:14 AM

 The Rise and Fall of Income Inequality in Mexico

Lustig, N. (1998). Mexico: The Remaking of an Economy (2nd ed.). Washington,

DC: The Brookings Institution.
Lustig, N. (2010). ‘El Impacto de 25 Años de Reformas sobre la Pobreza y la
Desigualdad’. In N. Lustig (ed.), Crecimiento Económico y Equidad. Mexico City: El
Colegio de México.
Lustig, N. (2011). ‘Scholars who Became Practitioners: The Influence of Research on
the Design, Evaluation and Political Survival of Mexico’s Anti-Poverty Program
Progresa/Oportunidades’. Tulane Economics Department Working Paper 1123. New
Orleans: Tulane University.
Lustig, N. and others (coordinator) (2011). ‘Fiscal Policy and Income Redistribution
in Latin America: Challenging the Conventional Wisdom’. Background paper for
Corporacion Andina de Fomento (CAF) Fiscal Policy for Development: Improving
the Nexus between Revenues and Spending (Política Fiscal para el
Desarrollo: Mejorando la Conexión entre Ingresos y Gastos).
Machado, J. A. F., and J. Mata (2005). ‘Counterfactual Decomposition of Changes
in Wage Distributions Using Quantile Regression’. Journal of Applied Econometrics,
20(4): 445–65.
Manacorda, M., C. Sánchez-Páramo, and N. Schady (2010). ‘Changes in Returns to
Education in Latin America: The Role of Demand and Supply of Skills’. Industrial
and Labour Relations Review, 63(2): 307–26.
Robertson, R. (2007). ‘Trade and Wages: Two Puzzles from Mexico’. The World
Economy, 30(9): 1378–98.
Scott, J. (2009). ‘Redistributive Constraints under High Inequality: The Case of
Mexico’. Paper prepared for the UNDP project ‘Markets, the State, and the
Dynamics of Inequality: How to Advance Inclusive Growth’.

163

oxfordhb-9780198701804-007.indd 163 9/27/2013 7:36:14 AM

View publication stats