Field: Labor Economics

Labor Economics

Researchers in the field of labor economics are fortunate to have available an abundance of high quality data with which to investigate research questions about the behavior of individuals and firms in labor markets. Perhaps most noteworthy are the large panel data sets on individuals, such as the National Longitudinal Surveys (NLS) and the Panel Study of Income Dynamics (PSID), which are utilized intensively by researchers, as well as the Current Population Survey (CPS). Newly available longitudinal data on firms from the Census Bureau will provide opportunities for interesting research on the behavior of firms in the labor market.
The goal of the labor economics program at UNC is to train students to conduct policy-relevant research on important questions in the field. This involves using theoretical models of individuals, firms, and markets as a basis for developing econometric models that can be empirically implemented, for example using data sets like those described above. We feel that policy-relevant research questions can best be addressed by combining models derived from economic theory with careful analysis of the data at our disposal. Therefore, students will be expected to be familiar with modern theoretical models of the labor market, applied econometrics, and to have substantial computer skills. Students should participate regularly in the Labor/Applied Microeconomics Workshop. To acquire skills in empirical analysis, which are difficult to transmit in a purely classroom context, students will be continuously involved in performing their own empirical research under the formal or informal direction of a faculty member in the field.
UNC Faculty
David Blau

David Blau is Professor of Economics and Fellow of the Carolina Population Center at UNC. He is a labor economist with research interests in the areas of child care, retirement, and wage inequality.

David Blau's vita

Brief descriptions of recent research papers

Tom Mroz

Thomas A. Mroz is a Professor in the Department of Economics, a Fellow in the Carolina Population Center, and a member of the Steering Committee for the Institute for Research in Social Science at the University of North Carolina, Chapel Hill. (Click here to visit his web page). He is a labor economist with interests in economic demography, health economics, transition economics, and applied econometrics. His research focuses on the estimation of behavioral responses of individuals to changes in their social and economic environments. His research ranges from studies of the determinants of fertility in France in the 18^th century, to the estimation of the consequences of illicit drug use on hours of work, to measuring the impacts of macroeconomic factors on wages and inequality over time, to assessing the determinants of women’s hours of work, and to evaluating the impact of social programs when programs might be targeted towards particularly needful groups. Much of his work focuses on data quality. He carried out an extensive evaluation of data quality and attrition biases in the National Longitudinal Survey of Youth, and he helped to establish, develop, and run the Russia Longitudinal Monitoring Survey (RLMS), the first nationally representative sample ever collected in the Russian Federation. Individual level data from 1998 will soon be available to researchers from the world wide web; researchers can easily link this data to earlier rounds of the RLMS.
Much of his work focuses on assessing the reliability of empirical research tools to yield unbiased and accurate estimates. He has evaluated nonparametric kernel regression procedures and has carried out extensive Monte Carlo studies of a class of semiparametric estimators to control for endogeneity and sample selection biases in complex empirical models. Current research projects include the estimation of the impacts of union membership on wages, the modeling of endogenous educational attainments in studies of fertility control in developing countries, and changes in gender inequality in Russia since the social and economic reforms of 1992. His current applied econometric work includes nonparametric conditional density estimation for studying the determinants of health care expenditures, and approaches to minimize the need to make out of sample projections in structural dynamic models.
Wilbert Van der Klaauw
Wilbert Van der Klaauw is an Associate Professor in the Department of Economics and a Research Fellow at the Carolina Population Center. His general areas of research are labor economics and applied econometrics, with a particular interest in the study of life cycle labor supply and occupational choice decisions, the economic determinants of household formation and dissolution, and the study of educational investment and productivity.(home page)

A unifying theme for much of this research is the empirical analysis of individual choice behavior over time through the estimation of structural dynamic models. This reflects a belief that such models are better able to explain observed life cycle choices and provide a richer basis for answering policy questions. His current work in this area deals with the joint analysis of savings and retirement decisions, and their dependence on earnings, pensions and social security benefits. Another component of this research project is the development of a methodology for the analysis and incorporation of self-reported intentions and expectations data into the estimation of dynamic behavioral models.
A second and more recent area of his research has been the development and application of alternative estimation strategies to evaluate the impact of social programs (or of interventions more generally defined), on individual choice behavior and outcomes. The focus in this research has been on providing reliable estimates to answer several more narrowly defined policy questions, such as: "What is the effect of serving in the military draft on future earnings?", "What is the effect of financial aid offers on college enrollment?", "What is the effect of the EEOC anti-discrimination law on minority employment?", and "What is the effect of Title 1 compensatory education programs on educational outcomes of disadvantaged students?". One approach that has received considerable attention in this research is a new powerful econometric evaluation design, called the Regression-Discontinuity Design.

Graduate Students
Recent graduates from the Ph.D. program who specialized in Labor Economics include:
Alison Auginbaugh, Ph.D., 1998. Economist, Bureau of Labor Statistics
Auginbaugh’s dissertation was supported by a Sloan Foundation dissertation fellowship, an honor reserved for the two dozen or so top economics graduate students in the country. She analyzes the relationship between the amount of child support payments received by single mothers and the achievement and ability of their children. This is an important issue because of the very rapid growth in the number of single-parent families and the increased government involvement in enforcing child support award collections. Previous research shows that the effect of child support income on child outcomes is larger than the effect of income from other sources. Alison develops a theory to explain this fact and provides an empirical analysis that allows her to test her theory. The empirical analysis uses data from the National Longitudinal Survey of Youth (NLSY) to estimate the relationship between child support income and child outcomes. Her empirical results show that lagged child outcomes have a strong positive impact on the amount of child support income received by the mother.
Charles Baum, Ph.D., 1999. Assistant professor of Economics, Middle Tennessee State University.
Baum’s dissertation examines how maternity leave mandates affect the labor market. He exploits the fact that during the late 1980s and early 1990s a dozen states passed legislation requiring firms to provide varying amounts of unpaid maternity leave, followed in 1993 by passage of the Federal Family and Medical Leave Act (FMLA), which guaranteed 12 weeks of unpaid leave. This created variation over time within and across states in the amount of leave firms are required to provide that can be used to identify the effects of such mandates. An especially attractive feature of his research design is that he can do before/after comparisons within states. This allows him to avoid the possible bias that would arise from relying solely on cross-state variation if there are unobserved factors associated with both passage of a maternity leave mandate in a state and features of a state’s labor market. Charlie uses data from the National Longitudinal Survey of Youth (NLSY) that provide a weekly record of the employment status, leave status, wages, and hours worked of a large sample of women who gave birth during the 1987-1993 period. Comparing the labor force outcomes of women who are covered by varying maternity leave mandates (including no mandate), controlling for many observable individual characteristics as well as state and year fixed effects, makes it possible to infer the effects of the amount of leave on the amount of time spent on leave, the timing of return to work, whether a woman returns to the job held before birth or a new job, and the wage rate and weekly hours of work at the time of the return to work. This provides a comprehensive picture of the effects of leave mandates.
Alexander Cowell, Ph.D., 1999, Research Scientist, Research Triangle Institute
Cowell’s dissertation examines the reasons for the well known empirical finding that higher educated individuals engage in healthier behaviors. The literature suggests two distinct theoretical explanations for the negative relationship between schooling and smoking, but it contains no practical guidelines for distinguishing between the competing explanations. Understanding the reasons for this relationship could help determine policy interventions that may lead to healthier behavior. Cowell’s dissertation develops a testable hypothesis that isolates the reason for the estimated relationship. The competing explanations operate through an efficiency mechanism and a future opportunity costs mechanism. To develop a testable hypothesis, he makes use of degree effects - the phenomenon of a discontinuous jump in earnings once a person gets a degree. Future opportunity costs will lead to degree effects appearing in the health behavior equations, while efficiency mechanisms would be unlikely to be manifested in degree effects. Using data from the National Longitudinal Survey of Youth, he find that detailed semi-parametric controls for the endogeneity of schooling actually eliminate much of the effect of education on binge drinking. For drinking, then, the education effects are spurious. The effect of education on smoking, however, remains after controlling for endogeneity, and his empirical analysis indicates that degree effects are important. He uses simulations to examine the effects of government policies that could operate on the health behaviors through future opportunity costs. His simulations indicate that the policies he considers do little to deter binge drinking. Moreover, despite the evidence that future opportunity costs deter smoking, only the extreme policy of forcing high school completion seems to have any discernable impact on smoking. By comparison, raising cigarette taxes seems to have a relatively large effect on deterring both smoking and binge drinking.
Elena Glinskaya, Ph,D., 1999, Young Professional, The World Bank
Glinskaya’s dissertation analyzes the determinants of job mobility of prime-aged Russian men using three waves of the Russia Longitudinal Monitoring Survey (RLMS) data collected in 1994, 1995, and 1996. The theoretical model she develops is consistent with matching, search, and human capital theories of job mobility. While the model predicts that jobs which pay relatively high wages are less likely to end, the model shows two competing effects of wage arrears on the probability of employment transitions. On one hand, an individual who is currently owed back wages might be more likely to change jobs than a peer without wage arrears, if that individual perceives current wage arrears as a signal of future non-payment. Alternatively, overdue wages might lock a worker into the enterprise. If the worker leaves the firm, he is less likely to receive any back wages, since an outsider has fewer instruments with which to press the management for payment. The empirical analysis demonstrates that, on average, an increase in the stock of wage arrears increases the probability of separation from the job. Further, jobs which withhold the total amount of the most recent monthly paycheck are more likely to end. This research also demonstrates that wages are an important determinant of job separation decisions for prime-aged Russian men. Jobs which pay higher wages are less likely to end. These findings show that current financial characteristics of jobs -- wages and wage arrears -- provide strong signals about future on-the-job financial outcomes.
Michael Lokshin, Ph.D., 1999, Economist, Development Economics Group, The World Bank
In his dissertation, Lokshin uses data from the Russian Longitudinal Monitoring Survey to estimate a model explaining the labor supply and child care demand decisions of mothers of young children in Russia. As in the U.S., there is both “formal” child care, in which a family pays a day care center or other provider to care for a child while the mother works; and informal child care, in which a family member or other relative cares for a child while the mother works. Informal care typically has little or no monetary cost, but requires the caregiver to sacrifice leisure or labor supply. The two forms of child care may also have different effects on child well-being. Hence the family must trade off income from the mother’s earnings against the monetary (caregiver’s earnings loss) or nonmonetary (caregiver’s leisure loss) cost of child care in deciding whether and how much the mother will work; and must trade off the monetary cost and child-well being benefit of formal care against the monetary or nonmonetary cost and child well-being benefit of informal care in deciding on the form of child care and the potential caregiver’s labor supply. Lokshin’s model incorporates these considerations and develops a theoretical model with a number of intuitive and testable predictions: (1) mothers with higher wage offers are more likely to work; (2) mother’s living in districts with a higher price of formal child care are less likely to work, and conditional on working are less likely to use formal care; and (3) families in which the potential informal child care providers have higher wages are more likely to have the potential providers work and are more likely to use formal care, conditional on the mother being employed. The empirical model is based directly on the theory, and allows him to test these hypotheses with few additional assumptions. The empirical results are consistent with the underlying theory: all of the hypotheses described are confirmed.
Timothy Savage, Ph.D., 1999. Economist, Charles River Associates.

Savage’s dissertation studies the medium to long run consequences youth unemployment. He develops a theoretical model of on-the-job human capital acquisition and considers an unemployment event as a shock to both the individual’s earnings and to the individual’s planned investments in training. As a consequence of this shock, the individual is forced off the planned human capital trajectory. A simple dynamic model indicates that individuals who recently experienced unemployment would have lower wages in subsequent years because of this negative shock to planned training, and they would undertake additional human capital investments in response to this shock to “catch-up” partially to where they would have been had they not suffered unemployment. Because of this optimal catch-up response to the negative unemployment shock, over time the wages of those experiencing unemployment would tend to converge to those of individuals who had not suffered the unemployment shock. Using data from the National Longitudinal Survey of Youth, he finds that each of these three theoretical predictions has substantial empirical support. Individuals do appear to undertake more training and work more hours after becoming unemployed. Wages do fall substantially after experiencing unemployment, with a six month unemployment spell being about the equivalent of losing a quarter of a year of education. As predicted by the theoretical model, however, the effect of unemployment on wages diminishes to almost zero within five years. The empirical analysis provides strong support for the theoretical characterization of youth unemployment as a temporary shock to human capital acquisition that is partially alleviated by optimal responses by the youth.

Erdal Tekin, Ph.D., 2001. Assistant Professor of Economics, Georgia State University

Tekin’s dissertation is an analysis of the impact of child care subsidies on labor force participation and enrollment in welfare programs in the U.S. This is a particularly important issue in the new post-reform welfare environment because the states now have a great deal of flexibility in using funds for child care subsidies and designing subsidy programs. He uses data from the National Survey of American Families, which are the only available data with information on child care subsidies. His study exploits variation across states in eligibility for and generosity of child care subsidies to determine the impact of such subsidies on the important welfare reform goals of increased labor force participation and reduced welfare dependence.

The results show that a lower child care price is associated with a higher probability that a single mother will be employed and pay for child care, and a lower probability that she will receive welfare benefits. The estimated effects of the price of child care are small, however, indicating that child care subsidies alone will be unlikely to generate large reductions in welfare caseloads. Other results indicate that the impact of the wage rate on employment and welfare participation is much larger than the impact of the price of child care.

Students currently working on dissertations in Labor Economics include:

Jeremy Bray
In his dissertation, Bray develops a theoretical model that links alcohol use to schooling and wages via a human capital production function. He shows that the effects of alcohol use on human capital formation cannot be isolated from simple substitution and income effects with a schooling demand equation. Instead, they can be estimated using a modified wage equation that allows for differing returns to education in years in which alcohol was used. He plans to estimate this modified wage equation using longitudinal data and state-of-the-art econometric techniques to control for both endogeneity and sample selection simultaneously.
Melissa Banzhaf

Banzhaf estimates a structural, dynamic model to examine how job turnover and wage growth differ by sex and education level. Understanding the job turnover and wage growth process is important because wages and wage growth are key measures of economic well-being. There are substantial differences in these key measures across men and women and across different education levels. Identifying the determinants of these differences is important both for general knowledge and for assessing potential roles for policy intervention. This study focuses on one potential determinant of the difference in wages between men and women, job turnover decisions, and seeks to quantify the impact of job turnover on wage growth. The data used in estimating the model are from the National Longitudinal Survey of Youth (NLSY) 1979 cohort. Banzhaf uses the NLSY data to construct a quarterly job history of each individual in order to estimate the parameters of the structural model. The model is estimated separately on four groups of individuals: low-educated men, high-educated men, low-educated women, and high-educated women, where education is defined by whether the individual has more than a high school degree or not. The parameters are then compared across groups and tested for equality.