Do Women Have Different Labor Supply Behaviors? Evidence Based on Educational Groups in Uruguay

ABSTRACT

This study uses Uruguay’s historical fluctuation in real wages to set up a natural experiment to measure the relation between women’s labor supply and wages. Using data from the Continuous Household Surveys of the Uruguayan National Statistics Institute, for 1986–2010, it aims to identify and explain heterogeneity in the labor supply behavior of women with different educational backgrounds. It finds that all women groups seek to optimize their remunerated work allocation throughout their life cycle, although women’s labor behaviors vary depending on educational levels. The rising trend of women’s labor force participation is expected to continue; its implications at the intensive margin are ambiguous and depend on how women conciliate paid work with household responsibilities – especially women with less education. These results could inform present debates about designing public policies to facilitate women’s entry into the labor market and cater to their different wage profiles and household care demands.

KEYWORDS:

• Labor force participation
• labor supply
• women and work

JEL Codes:

• C1
• J22

Acknowledgments

This research was supported by grants from Comisión Sectorial de Investigación Científica (Scientific Research Commission) from the Universidad de la República, Uruguay. The authors received helpful comments from the editor and three anonymous referees.

Notes

1 Because they are poorer and have lower income effect.

2 Marginal utility refers to the change in individual welfare derived from one additional unit of wealth.

3 James J. Heckman and Thomas MaCurdy (1980) develop a model based on panel data to estimate intertemporal elasticities; Richard Blundell, Costas Meghir and Pedro Neves (1993) estimate intertemporal elasticities for female supply and compare women with children to those without children; Pencavel (1998) distinguishes between married and single women; Donald Robbins, Daniel Salinas and Araceli Manco (2009) find little intertemporal elasticity – and even elasticity close to zero.

4 The World Bank (2015) ranked Uruguay as a high-income country with a gross national income per capita of US$16,810 in 2014.

5 According to the more recent figures available in the World Bank database, the fertility rate was 2.0 for Uruguay, 1.7 for OECD countries, and 1.9 for upper-middle-income countries, while the life expectancy at birth is 80.6, 82.7, and 76.3, respectively.

6 55.6 and 0.73 percent in Uruguay, and 50.9 and 0.74 percent in OECD countries.

7 Despite the limitation of this model assuming a “unitary” conceptualization of household labor decisions (see for example bargaining models and intrahousehold dynamics from a gender perspective in Agarwal [1997]). Nonetheless, when incorporating decisions throughout the life cycle, the intermissions in remunerated work participation that characterize the secondary worker’s decisions can be captured by the analysis.

8 There are three advantages to using pseudo-panels: they enable us to approach the intertemporal distribution of hours worked more closely (which is not possible with cross-section information); they can minimize bias due to errors in variable measurement at the micro-data level because they use averages or similar statistics by cohort; and they allow us to incorporate macroeconomic variables (Heckman and MaCurdy 1980; Pencavel 2002). In addition, pseudo-panels do not suffer from the problem of attrition (Deaton 1985; Antman and McKenzie 2005).

9 Bound, Jaeger, and Baker (1995) point out that the F statistical value below 10 would confirm the weakness of the instrument and the bias problems.

10 Pencavel (1998, 2002) justifies the use of external trade variables as an instrument by citing evidence found in previous studies in the US that indicate a link between these variables and the structure and level of wages.

11 Uruguay is a small open economy and as such is an international price taker, but the real exchange rate reflects the country’s overall global competitiveness as well as its monetary policy and exchange rate policy.

12 Alma Espino et al. (2012) included the projected job ratios as selection term in the estimation of women elasticities. However, their results do not have significant differences with and without this selection term.

13 Wages and incomes are deflated by the price consumption index, so they are expressed in real terms.

14 The definition of educational ranges is based on years of schooling completed: up to six years; between seven and nine years; between ten and twelve years of schooling; and more than twelve years.

15 The percentage of people in each cohort by educational level shows that the levels reached at 25 years old remained almost unchanged. Espino et al. (2014b) show the number of observations per cohort over time.

16 The basics of this decision are in Espino et al. (2014a, 2014b) where the construction criteria for the variable used in the estimations is also summarized.

17 It must be taken into account that women with more education tend to be overrepresented in sectors and occupations that involve fewer working hours. In this sense, women’s participation in the public sector and in health and education activities is relatively high. In 2010, the average hours worked in feminized sectors were 35.7 (social and health services), 32.9 (other community, social, and personal services), and 29.3 (education); other sectors with male predominance have more than 40 hours worked on average (INE 2010b).

18 Lifetime wages will continuously increase, with a steeper or flatter slope; however, in the short term, changes in real wages can be very pronounced depending on inflation.

19 The presence of children in the home increases women’s reservation wage, which lessens the probability that they will participate in the labor market. The costs involved in childcare depend on the availability of care options including whether relatives or friends are willing to provide care at a low or no cost (Deutsch 1998).

20 Considering that significant differences exist among women who compose the group without tertiary education (the dichotomous variables for educational ranges 2 and 3 are significant), we performed the regression for this group both with and without controls for years of education.

21 Among these, the percentage of women who are married or have a partner at the cell level has a positive impact, but we are unable to give an immediate interpretation of this effect. However, this last result is not robust, and in general the estimations of this coefficient are not significantly different from zero.

22 These fixed or cohort effects result from the estimation Equation 1(1) using IVA, without distinguishing by education groups. That is to say, it shows an average change between generations.

23 Jill Rubery et al. (2001) study the need to implement public policies that are compatible with gender equality in the market for the case of Europe.

24 The law is based on at least four arguments that are related to the needs of economic development: improvement of quality care, the lack of care services provisions due to greater demand of care, the decreasing number of people who are in a position to exercise care responsibilities, and finally, from a normative view, the gender equality and economic autonomy of women.

Additional information

Notes on contributors

Alma Espino

Alma Espino is an economist. She is Researcher at the Universidad de la República (Uruguay), and Professor of Gender and Economics in that Faculty. She is Department Coordinator of the Development and Gender Area at the Interdisciplinary Centre for Development Studies, Uruguay (CIEDUR). A member of the National System of Researchers in Uruguay, she has written or coauthored journal articles as well as contributed to several books. Her main research field is gender and labor economics.

Fernando Isabella

Fernando Isabella holds an MA and BA in Economics from the Universidad de la República, Uruguay. He is Researcher at the Instituto de Economía (Facultad de Ciencias Económicas y de Administración, Universidad de la República, Uruguay). His fields of research are economic development and labor economics.

Martin Leites

Martin Leites is a PhD student in the Department of Applied Economics at the Universidad Autónoma de Barcelona. He earned a BA in Economics from the Universidad de la República, Uruguay and an MA in Applied Economics from the Universidad Autónoma de Barcelona. He is full-time researcher at Instituto de Economía (Facultad de Ciencias Económicas y de Administración, Universidad de la República, Uruguay) and a member of the National System of Researchers in Uruguay. His research fields are: labor economics, income inequality, and income mobility.

Alina Machado

Alina Machado is a PhD candidate in Economics at Universidad Nacional de La Plata, Argentina. She earned an MA and a BA in Economics from the Universidad de la República, Uruguay. She is full-time researcher at the Instituto de Economía (Facultad de Ciencias Económicas y de Administración, Universidad de la República, Uruguay) and a member of the National System of Researchers in Uruguay. Her main fields of research are the economics of labor and education.