Ethnic wage gaps in Peru: What drives the particular disadvantage of indigenous women?

• DOI: http://doi.org/10.1111/j.1564-913X.2014.00016.x
• Date: 01 December 2015
• Author: Alexandre KOLEV,Pablo SUÁREZ ROBLES
• Published date: 01 December 2015

International Labour Review, Vol. 154 (2015), No. 4

Copyright © The authors 2015

Journal compilation © International Labour Organization 2015

* ERUDITE (Research team on the use of individual temporal data), Faculty of Economic

Sciences and Management, University Paris-Est Créteil Val de Marne, and International Training

Centre of the International Labour Organization (ITC-ILO), Turin, Italy, emails: kolev@u-pec.fr

and pablo.suarezrobles@gmail.com (corresponding author).

Responsibility for opinions expressed in signed articles rests solely with their authors, and

publication does not constitute an endorsement by the ILO.

Ethnic wage gaps in Peru:

What drives the particular disadvantage

of indigenous women?

Alexandre KOLEV* and Pablo SUÁREZ ROBLES*

Abstract. Ethnic inequality in labour market outcomes has long been on the pol-

icy agenda across Latin America. Using National Household Survey data, the au-

thors analyse Peru’s ethnic wage gap by sex over the period 2 005–11. Although the

gender wage gap is smaller than the ethnic gap, indigenous women compound both

disadvantages. Interpreting the unexplained component of the gap as an upper-

limit measure of ethnic discrimination, the authors then decompose the explained

component into the discrete effects of individual, job and locational characteris-

tics. Since more of the ethnic wage gap can be explained on these grounds among

women, they conclude with targeted policy proposals.

Ethnic gaps in employment, wages and career progression are a matter of

great concern in Latin America, one of the most unequal regions in the

world. In Peru, widespread ethnic inequality in labour market outcomes has

been on the national policy agenda for decades. The country has ratied the

ILO’s Indigenous and Tribal Peoples Convention, 1989 (No. 169), and since

2001, it has taken active measures aimed at integrating indigenous people in

local elections, public institutions and the central government, and in some

social programmes tackling poverty and exclusion; at the local and national

levels, it has also legislated penalties for discrimination (Sulmont Haak, 2010).

Yet, despite the Government’s efforts to protect ethnic minorities, indigenous

people – and indigenous women in particular – remain among the most vul-

nerable groups in the labour market, especially when it comes to wages.

From a policy perspective, it is important to explore the extent to which

ethnic wage gaps can be attributed to differences in observable characteristics,

such as human capital and job characteristics, or to discrimination. Evidence of

International Labour Review418

discrimination in pay can be derived from wage data decomposition analysis.

Specically, the share of the ethnic wage gap that is not explained by observ-

able characteristics (i.e. the unobserved component) is a ceiling estimate of

what can result from wage discrimination. If the wages earned by indigenous

people fall short of the wages earned by the ethnic majority group, after ac-

counting for differences in human capital and job characteristics, then there

is a strong case for supporting the enforcement of equal pay measures, along

with other interventions aimed at addressing ethnic disparities in educational

attainment and ethnic segregation in labour markets.

Economic theory tells us that there are two lead causes of discrimination.

“Taste-based” discrimination, which was rst modelled by Becker (1957), refers

to a situation where economic agents have a taste or a preference for some

persons and prefer not to interact with particular groups of people. Such dis-

crimination is costly but can nonetheless persist in competitive markets because

some individuals are willing to pay to avoid interacting with particular groups.

Another form of discrimination is based on imperfect information about work-

ers’ abilities – so-called statistical discrimination. According to this theory, which

was pioneered by Arrow (1973) and Phelps (1972), sex or ethnicity is believed

to be connected with workers’ productivity and thus explains a different treat-

ment of some particular groups. Within this theoretical framework, decomposing

the wage gap based on ethnicity can provide information on the extent to which

wage discrimination against particular ethnic groups is an issue.

Few studies have empirically analysed ethnic wage gaps in Latin Amer-

ica, so little is known about the factors that drive ethnic inequality in wages, the

changing patterns of such inequality and their determinants over time. This is

partly due to the absence of information on ethnicity in many household sur-

veys and national censuses. Yet, existing studies suggest that the ethnic wage

gap in Latin America is large, persistent, and wider than the gender wage gap.

For instance, in their analysis of seven countries with data on ethnicity (Bolivia,

Brazil, Chile, Ecuador, Guatemala, Peru and Paraguay), Atal, Ñopo and Winder

(2009) nd that the average wage premium for men compared to women stands

at 10 per cent, but rises to nearly 38 per cent for non-minorities compared to

minorities. They also establish that disparities in education and occupational

segregation account for the largest portion of this ethnic wage gap. On average,

however, one-third of the gap remained unexplained across the seven countries.

Beyond the ethnic wage gap, even less is known about the links between

gender and ethnicity. Indeed, the gender dimension of ethnic wage gaps is

often ignored by researchers exploring the sources of these wage differen-

tials in Latin America. The few studies available point to signicant gender

disparities within ethnic groups. In an analysis of four countries (Bolivia, Bra-

zil, Guatemala and Guyana), Cunningham and Jacobsen (200 8) use Oaxaca-

Blinder decomposition to analyse the wage differentials between whites and

non-whites and also compare men and women by ethnicity. They nd that the

portion of the ethnic wage gap that can be attributed to differences in observed

characteristics differs between the sexes.

Ethnic and gender wage gaps in Peru 419

In the case of Peru, studies of ethnic inequality in labour market out-

comes typically nd that while there are large inter-ethnic wage differentials,

discrimination may only explain a tiny portion of wage inequality. For instance,

in her analysis of the links between poverty and ethnicity, Trivelli (2005) un-

dertakes an Oaxaca decomposition of the ethnic wage gap using data from

the 2001 National Household Survey (ENAHO). She reports a substantial

wage gap between indigenous and non-indigenous workers – about 49 per

cent – but nds that differences in human capital endowments account for

43 per cent of this gap. Barrn (2008) uses the 2003 ENAHO to analyse the

impact of exclusion and discrimination on income inequality. He also nds a

large ethnic wage gap, with the mean labour income of indigenous workers

standing at only 56 per cent of that of non-indigenous workers, but that most

of the gap is explained by disparities in observed characteristics. In a sub-

sequent study based on the 2006 ENAHO, Atal, Ñopo and Winder (2009)

estimate at 45 per cent the raw wage gap between indigenous and non-indi-

genous people. They nd that when accounting for differences in observable

characteristics, the wage gap falls to 14 per cent. In contrast, Ñopo, Saavedra

and Torero (200 7), using the 2000 Living Standard Measurement Survey and

an extension of Oaxaca-Blinder decomposition, nd that a substantial part of

the differentials between racial groups cannot be explained by differences in

individual characteristics. They conclude that such results may be consistent

with employer discrimination.

In this article, we use annual data from the ENAHO for the period

2005 –11 to study the possible role of ethnic discrimination in generating wage

differences between indigenous and non-indigenous workers. The article con-

tributes to the literature on ethnic inequality in three ways. First, we provide

new evidence on the magnitude of discrimination against indigenous workers

in Peru. Second, we explore the gender dimension of ethnic wage gaps. Third,

our analysis casts new light on the trends in the ethnic wage gap and the

changing nature of its determinants over the period 2005–11. One important

objective of this article is to uncover the factors driving the particular disad-

vantage of indigenous women in Peru, drawing attention to the way in which

ethnic segregation and discrimination may differ between men and women.

The remainder of this article is organized into four sections. The rst

introduces the data and addresses methodological issues. The second pre-

sents basic descriptive statistics. The third section discusses our results, and

the fourth section concludes.

Data and methodology

To study the possible role of ethnic discrimination in generating wage dif-

ferences between indigenous and non-indigenous workers in Peru, we use

annual cross-section data from the ENAHO for the period 2005–11. The

ENAHO has been conducted on a continuous basis since May 2003 and is

representative at the national level: it covers both urban and rural areas in