Labor Informality: Choice or Sign of Segmentation? A Quantile Regression Approach at the Regional Level for Colombia

• Date: 01 November 2017
• Published date: 01 November 2017
• DOI: http://doi.org/10.1111/rode.12317

Labor Informality: Choice or Sign of Segmentation?

A Quantile Regression Approach at the Regional

Level for Colombia

Gustavo A. Garc

ıa*

Abstract

The labor market in developing countries is remarkably heterogeneous, with a small productive formal

sector characterized by high wages and attractive employment conditions, and a large informal sector

characterized by low productivity and volatile wages. The informal sector is particularly diverse. In this

paper, we examine the heterogeneity of the informal sector at the regional level in Colombia. In general,

our findings suggest that both voluntary and involuntary informal employment co-exist by choice and as

a consequence of labor market segmentation. We also find that there are striking differences in labor

market characteristics across cities, particularly with respect to informal employment.

1. Introduction

One of the most notable features of developing countries is the remarkable

heterogeneity of their urban labor markets. It is common to observe a small

productive formal sector that offers attractive labor conditions and relatively high wages

co-existing with a large informal sector that uses unskilled labor, is characterized by low

earnings and productivity, and does not fully comply with established legal regulations

(Dickens and Lang, 1985; Maloney, 1999, 2004; J€

utting and De Laiglesia, 2009).

Nevertheless, there is a considerable variety of workers within the large informal sector.

Why is there such diversity in the informal sector? Are there different kinds of

informal workers—some who are informal by choice and others who have no

alternative form of employment? Is labor informality a choice or the result of labor

market segmentation?

The segmented labor market theory considers informality a survival alternative to

involuntary unemployment for workers who are disadvantaged or rationed out of

formal employment opportunities (Dickens and Lang, 1985). The result is an

earnings dualism for individuals with similar characteristics that depend on the sector

in which they work. In the formal sector, there are internal markets that constrain

the labor supply and produce high wages, whereas the informal sector has no

institutional or efficiency–wage basis to regulate wages. In addition, low entry barriers

and an abundant supply of unskilled workers lead to low wages. Thus, wages depend on

the sector in which a worker is employed and not on his or her skills per se (Uribe

et al., 2 007).

In contrast, the orthodox neoclassical view of the human capital theory postulates

that, as in any another market, price flexibility and free labor mobility lead to a full

*Garc

ıa: Research in Spatial Economics (RiSE-group) and Department of Economics, Universidad

EAFIT, Medell

ın, Colombia. E-mail: ggarci24@eafit.edu.co. The author wishes to thank participants at

the 16th IZA European Summer School in Labor Economics and thank Jose

´Luis Roig, Josep Lluı

´s

Raymond, the editor and two anonymous referees for their suggestions and comments.

Review of Development Economics, 21(4), 985–1017, 2017

DOI:10.1111/rode.12317

©2017 John Wiley & Sons Ltd

employment equilibrium with equal remuneration for the same kind of work (De

Soto, 1987; Saavedra and Chong, 1999; Maloney, 1999). Given this competitive

market framework, participation in the informal sector may be a desirable choice for

workers and firms, based on private cost–benefit calculations of inclusion in the

sector. Informality can have desirable non-wage features and thus individuals

maximize their utility rather than their earnings. Alternatively, certain workers have

a comparative advantage in the informal sector that does not exist in the formal

sector (Gindling, 1991).

These two polarized views can be combined if the informal sector is very

heterogeneous and contains elements of each scenario; that is, if the informal sector

has its own internal duality. The literature has recognized the existence of “upper”

and “lower” tiers or “voluntary” and “involuntary” entry of informal employees or

firms (Fields, 1990, 2005; Cunningham and Maloney, 2001; Maloney, 2004). In this

scenario, upper-tier employees are those who are voluntarily informal because,

based on their specific characteristics, they expect to earn more in the informal

sector than they would in the formal sector. In contrast, lower-tier employees are

disadvantaged workers who see informality as a last resort.

From an empirical stance, this notion of dualism within the informal sector has

not been satisfactorily addressed. For example, Magnac (1991), who tested the

labor market of Colombia in the 1980s for competitiveness or segmentation, found

evidence of a competitive labor market structure. Similarly, Gindling (1991) and

Pratap and Quintin (2006) found evidence of segmentation in Costa Rica but a

competitive structure in Argentina. However, in all of these studies, the authors

assume homogeneity of the informal sector, thereby limiting their analysis.

Among the few studies that have tried to model the heterogeneous structure of

the informal sector are Cunningham and Maloney (2001) and G€

unther and Launov

(2012). Cunningham and Maloney (2001) model the informal sector as a mixture of

“upper-tier” and “lower-tier” enterprises, using the econometric technique of factor

and cluster analysis to allow for segmentation of the market. They find evidence of

segmentation; however, they only consider informal firms and thus their model

excludes the possibility of being a formal firm. Further, the authors do not take into

account the selection bias induced by workers’ decisions about type of employment.

G€

unther and Launov (2012) analyze the possible heterogeneous structure of the

informal sector, estimating a finite mixture model that allows them to determine the

number and size of segments that might compose the informal sector. This model

uses minimal a priori assumptions to determine the segments and provides a new

method for identifying the extent of voluntary and/or involuntary employment in

the informal sector. The empirical analysis uses data from the Ivory Coast at the

end of the 1990s. Among their findings, the authors report that the informal sector

comprises two segments: a highly paid segment and a low-paid segment. They also

found that 45% of informal employment is involuntary and mainly located in the

low-paid informal segment, whereas the remaining 55% of informal employment is

voluntary and situated in the highly paid informal segment.

This paper provides new evidence on the heterogeneity of the informal sector in

Colombia. The novelty of this study is twofold. First, we analyze the informal sector in

Colombia, which provides rich evidence from a large and heterogeneous informal

sector that presents a particularly interesting case for analyzing heterogeneity in the

informal sector in developing countries. The extant literature reports mixed and

inconclusive results on the heterogeneity of the informal sector; this paper uses the

Colombian case to offer more evidence on the issue. Specifically, we analyze the

986 Gustavo A. Garc

ıa

©2017 John Wiley & Sons Ltd

factors that affect wage differentials between formal and informal workers using

quantile decomposition methodology, controlling for selectivity bias caused by

correlated unobserved heterogeneity affecting wages. We employ the approach

introduced by Albrecht et al. (2009) based on Buchinsky (1998), which is a non-

parametric method to account for selection in quantile regression analysis. This

methodology has two advantages. First, it allows us to analyze the heterogeneity of

the informal sector by considering a decomposition of the wage differential between

the formal and informal sectors throughout the entire distribution of wages. Second, it

is one of the few approaches empirically available in the literature for selectivity

correction in quantile regression on cross-sectional data.

Our second contribution is to conduct the analysis of formal–informal wage gap

decomposition at a regional level. The extant literature typically considers a national

context and does not take into account the possibility of regional heterogeneities

among informal workers. Given the geographic, demographic and social conditions

and the economic dynamics of Colombia, there are marked differences in the

structures and dynamics of the local labor markets, which could imply that informality

is not homogenous throughout the territory. In Colombia, roughly six out of ten

employees work in the informal sector.

1

Cities such as C

ucuta or Monter

ıa have

informality rates of approximately 75%; others, including Medell

ın and Bogot

a, have

rates of approximately 50% (Garc

ıa, 2011; Galvis, 2012).

The decomposition of the wage differential between the formal and informal

sectors throughout the entire distribution of wages allows us to determine the

proportion of the wage gap that is due to differences in prices related to individual

characteristics and the proportion resulting from characteristics that differ between

the formal and informal sectors. If the wage gap is mainly attributable to the former

factor, it indicates that individuals in the informal sector earn less because they earn

lower returns on their skills and therefore are members of the disadvantaged sector of

a segmented market. In contrast, if the wage gap is primarily explained by the latter

factor, labor segmentation is not as significant as it is in the former case and wage

differences between sectors are due to differences in endowments. In the latter

situation, being an informal worker is a choice, because these individuals can obtain

non-wage benefits or earn more than they would in the formal sector.

Following this introduction, section 2 summari zes the literature on wage

differences between formal and informal workers. Section 3 describes the data. In

section 4, we discuss the estimation procedure. Section 5 presents the empirical

findings, and conclusions are drawn in section 6.

2. Literature Review

Although the empirical evidence on wage differentials between the formal and

informal sectors in various countries is considerable, the results are mixed and

inconclusive. Most early studies analyzing the traditional segmented labor market

theory find that workers in the formal sector are better rewarded than their

counterparts in the informal sector. These studies include Mazumdar (1981) in

Malaysia, Heckman and Hotz (1986) in Panama, Roberts (1989) in Mexico, Marcouiller

et al. (1997) in El Salvador and Peru, Tansel (1999, 2000) in Turkey and Gong and Van

Soest (2002) in Mexico. Morerecently, Tansel et al. (2015) use a fixed effects model with

a quantile regression technique and find a persistent informal wage penalty in Egypt.

In contrast, other studies suggest that the formal–informal wage gap may not be

a stylized fact. For instance, using data for urban Bolivia, Pradhan and Van Soest

INFORMALITY: CHOICE OR SIGN OF SEGMENTATION? 987

©2017 John Wiley & Sons Ltd