Glass Ceilings versus Sticky Floors: Evidence from Southeast Asia and an International Update

• DOI: http://doi.org/10.1111/asej.12056
• Published date: 01 September 2015
• Date: 01 September 2015
• Author: Zheng Fang,Chris Sakellariou

Glass Ceilings versus Sticky Floors:

Evidence from Southeast Asia and

an International Update*

Zheng Fang and Chris Sakellariou

Received March 10, 2013; accepted May 17, 2015

Using counterfactual decompositions combined with quantile regression, we inves-

tigate the pattern of gender wage differentials in Asian and Latin American coun-

tries and combine the findings with existing evidence for European and other

mostly developed countries. While in the group of Latin American countries glass

ceilings are prevalent, no clear evidence of glass ceilings is found in the group of

Asian countries where, generally, sticky floors or a mixed pattern is the norm. The

findings are robust with as well as without occupation controls. In addition, analy-

sis by sector points to glass ceilings in the public sector in most countries, while in

the private sector the patterns vary. Combining the new evidence from the present

study with existing evidence (in total we consider 60 countries), a comprehensive

global picture of gender wage differentials is provided. Possible explanations for

differences in patterns of gender wage differentials are discussed.

Keywords: gender wage differentials, glass ceilings, quantile decomposition,

sticky floors.

JEL classification: J31, J16, J7.

doi: 10.1111/asej.12056

I. Introduction

The gender wage gap has been extensively studied at the mean; however, how it

evolves along the wage distribution has only received attention in recent years.

For example, Albrecht et al. (2003) found an increasing gender gap along the

wage distribution in Sweden. Arulampalam et al. (2007) studied 11 European

countries and found that the glass ceiling effect is prevalent. A glass ceiling is an

observed gender gap that is wider at the top of the counterfactual wage distribu-

tion compared to the bottom.1

In this paper, we use modern counterfactual decomposition techniques com-

bined with quantile regression estimation to explore gender wage differentials in

a group of Asian countries (for which evidence is lacking), and several countries

* Fang: Nanyang Technological University, HSS-04-75, 14 Nanyang Drive, Singapore 637332.

Email: fang0031@ntu.edu.sg. Sakellariou (corresponding author): Nanyang Technological Univer-

sity, HSS-04-64, 14 Nanyang Drive, Singapore 637332 Singapore. Email: acsake@ntu.edu.sg.

1 Correspondingly, a wider gap at the bottom of the counterfactual wage distribution is referred to

as a ‘sticky floor’.

bs_bs_banner

Asian Economic Journal 2015, Vol. 29 No. 3, 215–242 215

© 2015 The Authors

Asian Economic Journal © 2015 East Asian Economic Association and Wiley Publishing Pty Ltd

in Latin America, to expand the body of evidence in the glass ceilings/sticky

floors literature and provide an international update. We summarize the findings

from approximately 60 countries in 5 continents and suggest potential explana-

tions to the stylized facts documented across countries. The approach used is a

refinement of the popular Machado and Mata (2005) procedure proposed by

Melly (2006).

The counterfactual wage distribution in the present study is constructed using

women’s characteristics and men’s payoff structure; therefore, the difference

between the wage distribution of male workers and the counterfactual wage

distribution can be explained by gender-specific attributes, whereas the difference

between the counterfactual distribution and the distribution of female workers

measures the unexplained wage gap (coefficients component). In other words, it

is a proxy of the degree of ‘discrimination’ against women in the labor market.

Two alternative specifications are used: in the first, a vector of basic variables is

controlled for; subsequently, occupation dummies are added into the covariates to

assess the effect of occupation on the estimated counterfactual wage gap.2Con-

clusions concerning glass ceiling or sticky floor effects are drawn on the basis of

observed profiles of the counterfactual gender wage gap. When possible, the

analysis is redone by sector of employment (public vs private). As has been

widely recognized, wage setting institutions in the private sector are quite differ-

ent from that in the public sector (see e.g. Patrinos et al., 2009). Sample selection

is not considered in this study for two reasons: first, research on how to adjust for

sample selection within quantile regression is ongoing (Buchinsky, 2001; Huber

and Melly, 2011); and second, there is a lack of suitable instrumental variables in

most of the datasets.

The paper is organized as follows, Section II summarizes the origins of gender

wage differentials in the literature. Section III introduces the conceptual frame-

work and specifies the counterfactual decomposition technique employed in this

study. Data sources are described in Section IV and the results are presented and

discussed in Section V. The final section concludes.

II. Origin of Gender Wage Differentials

There are many potential explanations as to why seemingly identical men and

women earn different wages and why the gap varies across the wage distribution.

We summarize and distinguish explanations from the demand side as well as the

supply side of the labor market.

2 Albrechet et al. (2003) argue that occupation should not be viewed as a factor accounting for the

male–female wage gap, but, interestingly, they found a glass ceiling in Sweden even after occupation

is taken into considered. In this paper, we present two groups of regressions (with and without

controlling for occupation) to assess its impact on the findings of glass ceilings or sticky floors.

ASIAN ECONOMIC JOURNAL 216

© 2015 The Authors

Asian Economic Journal © 2015 East Asian Economic Association and Wiley Publishing Pty Ltd