Home Address:
  420 Temple Street, Room 419
  New Haven, CT 06511

Telephone: (203) 606-8389 (cell)
                   (203) 436-2113 (home)

Labor Economics
Public Economics
Applied Microeconomics

Microeconomics
Public Economics
Labor Economics
Applied Econometrics
Industrial Organization

(Oral) May 2001, Public Economics, October 2001, Industrial Organization
(Written) May 2000, Microeconomic and Macroeconomic Theory

"Female Labor Supply and Gender Gaps: Roles of Childbearing and Home Responsibility"

Professor Joseph Altonji
Professor Hanming Fang
Professor T. Paul Schultz

May 2006

M.Phil., Economics, Yale University, 2003
M.A., Economics, Yale University, 2001
B.A., Economics, Beijing University, Beijing, 1999

Yale University Dissertation Fellowship, 2003
Yale University Graduate Fellowship, 1999-2002

Teaching Fellow:
Introductory Microeconomics, Yale University, 2003, 2004, 2005
International Trade, Yale University, 2003
Intermediate Microeconomics, Yale University, 2002
Introductory Macroeconomics, Yale University, 2002
Introduction to Probability and Statistics, Yale University, 2001

Professor Joseph Altonji
Department of Economics
Yale University
PO Box 208264
New Haven, CT 06520-8264
Telephone: (203) 432-6285
Fax: (203) 432-6323
Email: joseph.altonji@yale.edu

Professor Hanming Fang
Department of Economics
Yale University
PO Box 208264
New Haven, CT 06520-208264
Telephone: (203) 432-3547
Fax: (203) 432-6323
Email: hanming.fang@yale.edu

My dissertation concerns two important subjects in labor economics: women’s labor supply and the gender gap. The U.S. economy witnessed a dramatic rise in women’s participation in the labor force, from 34 percent in 1950 to 60 percent in 1998, while the gender gap in pay showed only moderate and occasional progress – the women-to-men pay ratio remained 60 percent until 1980, rose to around 75 percent in the mid-1990s, and then leveled off. Within this broader context, my research explores the role of childbearing and home responsibility in women’s labor market outcomes and the gender gap.

In my first paper, Women’s Labor Market Outcomes and Childbearing (job market paper), I estimate the causal effect of having children on women’s labor supply and earnings. Previous studies have addressed the fact that women’s fertility is endogenous by using instrumental variables for fertility, such as having a twin birth or having the first two children of the same sex. (See Bronars and Grogger (1994) and Angrist and Evans (1998)). I use information from a woman’s pregnancy history as a source of exogenous variation in fertility. For example, in one specification I consider women who have had at least one completed pregnancy and use whether the first pregnancy ended in a miscarriage as an instrumental variable for whether the woman ultimately has one or more children. My instrumental variables can be constructed from the very detailed information on each woman’s pregnancy and fertility histories provided by the National Survey of Family Growth surveys, the data for my study. I show that a miscarriage has a significant negative impact on the number of children that a woman will ultimately have. The link to fertility may arise because childbearing years are limited and because a miscarriage might indicate that the woman has physiological problems with carrying a pregnancy to term.

My identification strategy permits me to estimate the effect of ever having a child on women’s labor market outcomes. One cannot do this using twin births or children of the same sex as instruments. Identifying the causal effect of having at least one child is of special importance, given the large gap in labor supply between childless women and women who have children.

The IV results show a small negative statistically significant impact of having children on a woman’s various labor market outcomes. For example, having children lowers a woman’s employment probability by about 5 percent, weekly hours of work by about 6 hours, and annual earnings by about 2700 dollars. The effects of having more than one child and having more than two children are similar in magnitude. In addition, consistent with Angrist and Evans (1998), I find that for almost all labor market outcomes the instrumental variables estimates are notably smaller in magnitude than the estimates obtained from ordinary least squares estimation, suggesting that OLS estimates overstate the causal effect of childbearing on women’s labor supply and earnings.

In the second paper, Home Responsibility, Employer Perceptions and Endogenous Gender Gaps (in progress), I construct a model in which a gender gap in the labor market can arise in equilibrium even if it is assumed that men and women have ex ante identical distributions of characteristics relevant to the labor market and to home production. That is, the gap can arise even if women do not have a comparative advantage at home. In particular, the model highlights the role of home responsibility in facilitating the possibility of a gender gap in the economy. The importance of home responsibility in studying the gender gap is suggested by research such as Hersch and Stratton (2002), who show that the gender gap is lowered once one controls for household work.

My model is related to Francois (1998) and Albanesi and Olivetti (2005) in that all three models incorporate two places in the economy, work and home, where the outcome in one affects the decision making in the other. Following Coate and Loury (1993), employers have imperfect information about a worker’s responsibilities at home, which affect the worker’s productivity on the job (e.g., a person who is the primary care taker at home may be more likely to disrupt his/her work because of a family emergency than a person who is not). Couples make optimal decisions regarding the home responsibility by taking into account both the abilities and the labor market opportunities of the husband and wife. The solution to the couple’s problem is found through a revealed preference argument, as in Fang (2004). In equilibrium, if employers expect that men on average are less likely to take on major home responsibilities than women, they will give men favorable treatment in the work place. In turn, this gives incentives to couples to make decisions such that indeed it is more likely for women rather than men to be the primary person at home, thus confirming the employer’s initial expectation.

The model yields testable predictions regarding the relationships among the work status, career achievements, earnings abilities, and home responsibilities of husbands and wives. I will investigate these relationships using data from the Current Population Survey and the Panel Study of Income Dynamics.