Volume 43, Number 171,
October-December 2012
Maquiladora Factories and Household
Income in Yucatan
Javier Becerril, Rafael Ortiz and Lilian Albornoz
ANALYTICAL TOOLS

The basic relationship that this study looks at to examine the impact of the decision to work for an mei is the well-being of individuals, measured through their income, as a linear function of a vector of explanatory variables (Xi) and a dichotomic decision variable (Ri). The linear regression equation is as follows:

(1)

Where Yi is the average income of individual i: the literal ui is the random error term with normal distribution and Ri is the dichotomic variable 1 or 0, regarding the decision to work for an mei or not. Ri = 1 if the individual works for the mei and Ri = 0 if the person works in another job. The vector Xi represents the individual’s characteristics, home and location of residence. The decision to work or not work at an mei is independent of the profiles of individuals and their homes, because an individual’s decision to work for an mei is based on his own decisions. In other words, it is not a random decision.

Assuming that individuals assume a neutral risk in the decision to work for the mei, then the function regarding the decision to work for an mei can be expressed as:

(2)

Where is a latent variable that denotes the difference between the utility of working for an mei, , and the utility or benefit of not working for an mei, . The individual will decide to work for an mei if . The term provides an estimate of the difference in the utility of working for the mei , using the profile of the individuals and the characteristics of their homes Xi, as explanatory variables, while is the error term. To estimate equations (1) and (2), it is necessary to describe the relationship that exists between being an mei worker and income level. These equations may be interdependent. In other words, being an employee of an mei could help increase household income and contribute to changing or not changing the condition of poverty (from poor household to not poor household). However, households that are not poor will probably have a more ad-hoc profile to work for an mei. In this way, assigning “treatment (individuals that work for an mei)” is not random. Within the group of individuals that work for an mei it will be systematically different. Particularly, there will be self-selection bias if unobservable factors influence both terms of error in the income equation (ui) and the choice to work or not to work (), and may result in correlation between the terms of disturbance in the income (1) and decision (2) equations. Thus, estimating equation (1) using the least squares (ls) method is not the correct approach because it would obtain biased and inefficient estimators.

Some others have used Heckman’s two-stage method to correct the bias, when the correlation of the two error terms is greater than zero. However, this approach depends on the restrictive assumption of normal distribution of errors. Another mechanism to control the self-selection bias is to use instrumental variables (iv). However, the biggest limitation of this method is that it is difficult to find and identify iv in the estimate. Moreover, both the ls and iv methods tend to impose a linear function, where the assumption implies that the coefficients over the variables of control of the individuals that work and do not work for mei are similar. As such, this assumption probably does not correct the bias, because the coefficients may differ (Jalan and Ravallion, 2003). Unlike the previously mentioned parametric methods, Propensity Score Matching (psm) does not require assumptions regarding the functional form to model a specific relationship between the results and the prediction of results. The disadvantage of the psm method is the Conditional Independence Assumption (cia), which says that for a given set of co-variables, participation is independent of potential results. Smith and Todd (2005) argue that there may be systematic differences between those who work and do not work for an mei, and they may even have an impact on observable variables. Jalan and Ravallion (2003) argue that assumptions on selecting observable variables are no more restrictive than those upon which the Heckman two-stage method is based or iv, when they are used in the analysis of cross-section data.

Published in Mexico, 2012-2017 © D.R. Universidad Nacional Autónoma de México (UNAM).
PROBLEMAS DEL DESARROLLO. REVISTA LATINOAMERICANA DE ECONOMÍA, Volume 49, Number 193, April-June 2018 is a quarterly publication by the Universidad Nacional Autónoma de México, Ciudad Universitaria, Coyoacán, CP 04510, México, D.F. by Instituto de Investigaciones Económicas, Circuito Mario de la Cueva, Ciudad Universitaria, Coyoacán,
CP 04510, México, D.F. Tel (52 55) 56 23 01 05 and (52 55) 56 24 23 39, fax (52 55) 56 23 00 97, www.probdes.iiec.unam.mx, revprode@unam.mx. Journal Editor: Moritz Cruz. Reservation of rights to exclusive use of the title: 04-2012-070613560300-203, ISSN: pending. Person responsible for the latest update of this issue: Minerva García, Circuito Maestro Mario de la Cueva s/n, Ciudad Universitaria, Coyoacán, CP 04510, México D.F., latest update: June 27th, 2018.
The opinions expressed by authors do not necessarily reflect those of the editor of the publication.
Permission to reproduce all or part of the published texts is granted provided the source is cited in full including the web address.
Credits | Contact

The online journal Problemas del Desarrollo. Revista Latinoamericana de Economía corresponds to the printed edition of the same title with ISSN 0301-7036