2012–2014 Emerging Scholars Research Grants:
Building Human Capital and Economic Potential

IRP awarded funding to five projects as part of the Building Human Capital and Economic Potential research initiative, with a maximum award of $20,000 each, to emerging scholars as defined below. The awards run from May 1, 2013, through August 2014. Throughout the award period, grantees will benefit from consultation with IRP senior affiliates, with each other, and—during a workshop at which grantees will present their draft paper—with other senior poverty scholars.

Funded Proposals

Proposal Abstracts

What Job Characteristics Do Mothers of Very Young Children Value the Most?
Principal Investigator: Flavio A. Cunha, University of Pennsylvania

Flavio A. Cunha

In real life, people reveal their preferences through choices. Understanding how changes in job attributes affect the decision to work or not is important for many researchers who are interested in studying poverty, welfare dependence, and the even accumulation of human capital. The central question this proposal aims to answer is: How much are mothers of young children willing to pay for the opportunity to have access to jobs that offer benefits such as child-care benefits and flexible working schedules that match the hours of operation of child care providers? In spite of the importance of these questions, little is known about these issues.

The goal of our research is to collect new and unique data that allows us to estimate the marginal willingness to pay and the elasticity of labor supply with respect to job attributes such as hourly wages, working hours (e.g., part time vs. full time), working schedule (e.g., flexible working hours vs. office working hours), child-care benefits (e.g., no benefits vs. employer-sponsored child-care vouchers), distance (measured in commuting time). Following the tradition in economics, one source of data information will be to ask subjects who are working to report the attributes of the jobs they have. We will build on the seminal work by James Heckman to account for selection in order to estimate the parameters that determine the labor supply of the mothers of young children. This is known in economics as the Revealed Preference Approach (RP).

In addition to this data, data on abstract choices regarding labor supply will also be collected. In order to do that, we follow the tradition in Marketing research and systematically construct choice alternatives following an experimental design. This is known as the Conjoint Analysis (CA), Discrete Choice Experimentation (DCE), or Choice-Based Conjoint Analysis (CBC). The advantage of this approach is that it allows us to bypass the selection problem: The information on abstract choices made by individuals can be collected for the subjects that are and are not working.

The RP and CA data will be collected by expanding the scope of the Philadelphia Human Development Study (PHD Study). The PHD Study is an NIH-sponsored longitudinal research project that aims to measure the determinants of child development. It will enroll 960 primiparous women in Philadelphia. The study's first round takes place when the subjects are in their second trimester of their first pregnancy, and is scheduled to start in the summer of 2013.

Specifically, we will:

  1. Estimate parameters that describe the maternal preferences for job attributes using the RP data and accounting for selection.
  2. Estimate parameters that describe the maternal preferences for job attributes using the CA data.
  3. Investigate if the different approaches provide similar or different answers.
  4. Estimate the marginal willingness to pay for job attributes such as child-care benefits, flexible working hours, and access to employer-sponsored health insurance.
  5. Estimate the labor force attachment of mothers of young children if jobs offered attributes that matched maternal preferences more closely.

Using the data from the proposal, we will test the following hypotheses:

H1: Maternal elasticity of labor supply with respect to a 10 percent increase in child-care benefits is much higher than the elasticity of labor supply with respect to a 10 percent increase in hourly wages.

H2: Maternal labor force attachment can be promoted by making cost-neutral changes to job attributes (e.g., increase the number of jobs that offer working schedule flexibility to mothers of young children).

H3: The parameters obtained from the RP data are statistically comparable to the ones obtained by the CA data.

The Role of Career and Technical Education in Promoting Human Capital Accumulation and Bridging Labor-Market Needs: Evidence from Massachusetts
Principal Investigator: Shaun Dougherty, University of Connecticut

Shaun Dougherty

The overall objective of my proposed work is to understand whether and how career and technical educational (CTE) programs in Massachusetts that engage in public-private partnerships promote human capital accumulation in youth (as measured by high-school graduation), and whether existing program offerings match demonstrated labor market needs, especially in areas with high concentrations of individuals from low-income families.

To undertake this analysis I propose to use administrative data from the Massachusetts Department of Elementary and Secondary Education as well as publicly available employment data from the Bureau of Labor Statistics to track employment trends within the state. I will estimate the effect of participating in a CTE program that utilizes public-private partnerships on high-school graduation using coarsened exact matching. Under this matching approach I propose to match treated students with similar control students based on observable criteria that are employed by CTE programs in Massachusetts to admit students to their schools. By creating these treatment-control comparisons I can get plausibly causal estimates of the impact of CTE participation, with an emphasis on understanding the impact of participation for lower-income students.

Understanding the impact of CTE participation on attainment of a high-school diploma will inform policymakers and educators as to the potential effects of CTE in promoting or hindering the human capital accumulation of youth in Massachusetts. My proposed descriptive work will emphasize whether and how these programs are positioned to support training in areas for which there is labor-market demand.

Minimum Wages and the Distribution of Family Incomes
Principal Investigator: Arindrajit Dube, University of Massachusetts–Amherst

Arindrajit Dube

Research Objectives: Extant studies of minimum wages on family incomes suffer from the focus on small and often disjoint sets of distributional outcomes and demographic groups, use of aggregate data, small samples and lack of precision, and lack of sufficient individual-level and geographic controls. The findings from these various papers are often at odds with each other. The proposed research will use individual level data from 1990 to 2011 and consider a large array of distributional measures, a wide range of cutoffs for low-income definitions, numerous demographic groups, and a rich set of controls to show the effects of minimum wage policies in the U.S. on family income distribution. Besides estimating the impact on poverty rates and poverty gap index for a wide range of thresholds, I will estimate unconditional quantile effects of minimum wages on the family income distribution. I will explore heterogeneities in impact by time periods and demographic groups, and interactions with EITC. The breadth of the outcome measures, and the extent of controls, and the length of the sample period make this analysis the most comprehensive evaluation of minimum wage policy on family economic well-being. Use of better methods, additional data, and an effort to rationalize the somewhat contradictory extant evidence can help provide valuable information to policymakers and the public at large about a timely topic given current discussions to raise and index the minimum wage.

Data: I will use individual level the March Current Population (CPS) survey between 1990 and 2011. I will augment the CPS data with data on state level EITC programs from the University of Kentucky Center for Poverty Research, and state and federal minimum wages from the Department of Labor. The core demographic groups will include: all individuals under 65, individuals with high school degree or less (HSL), HSL and between the ages of 18–45, and finally HSL between the ages 18–45 with kids. I will also show the key results for other demographic groups that have been considered in the literature. These include (1) children, (2) single mothers (3) workers, (4) young adults and (5) teens. Finally, I will also provide estimates by race and citizenship status.

Research Methods: I will estimate the impact of minimum wage increases on quantiles of the unconditional family income distribution, using recentered influence function (RIF) approach developed by Firpo, Fortin, and Lemieux (2009, 2010). The unconditional quantile partial effect (UQPE) tells us how a unit increase in minimum wage affects, say, the 25th quantile of the family income distribution, after accounting for other covariates. I will also estimate impacts on headcount poverty rates and poverty gaps for a wide range of income thresholds (between 25 percent and 350 percent of the federal income guidelines). The RIF approach also allows us to estimate the impact on poverty gap using individual level data and controls. I will use a rich array of individual characteristics, state characteristics, state fixed effects and region-specific time effects to control for unobserved heterogeneity. Robustness tests will include additional controls such as state-specific trends and recession controls. I will also consider robustness of results to construction of poverty measures and equivalence scales.

Using Mobile Technology to Improve Academic Performance and Persistence among Community College Students: An Experimental Evaluation of a Behavioral Intervention
Principal Investigators: Patrick Sharkey and Joshua M. Aronson, New York University

Patrick Sharkey

Objectives: This project will use a randomized block design to evaluate an intervention to increase rates of academic achievement and persistence for a cohort of entering students at a two-year community college in central Connecticut. The intervention combines well-developed theory from behavioral economics and social psychology with an innovative mobile platform that will provide students with consistent behavioral prompts, "nudges," and supports that have been shown to be effective in improving persistence and academic performance among low-income students and students of color.

Joshua M. Aronson

Data: The investigators have partnered with Persistence Plus, an educational technology firm that has contracted with Middlesex Community College to provide their entering cohort of students in the fall of 2013 with access to the services offered through their mobile platform. The investigators will have access to all administrative records on the entering cohort of students in the fall of 2013, which will be used to create the following measures of retention and academic performance: indicators of completion of all or any enrolled classes in the fall and spring semesters; indicators for enrollment in the spring semester; passing grades in all or any enrolled courses in the fall and spring semesters; grade point average in the fall and spring semesters.

Methods: Students in the entering cohort will be classified into blocks based on employment status, age/family structure categories, and gender. Within blocks, students will be randomly assigned to the treatment group, which will receive all services provided by Persistence Plus, or the control group. Analyses will estimate the effect of being offered access to the treatment, the effect of actively engaging with the mobile platform throughout the semester, and the possibility of treatment effect heterogeneity.

The Effect of School-to-Work Programs on School to Labor Market Transitions
Principal Investigators: Kevin Stange and Daniel Kreisman, University of Michigan

Kevin Strange

We ask what effect School-to-Work (STW) programs have on the transition from high school to college, on transitions from high school to the work force, and on job match, employment and earnings. Increases in the number of students who enroll in college and then drop out, and in the number of students who enter the labor force after high school only to later enroll in college, reflect increasing ex ante uncertainty in the college going decision. These "transitions" are costly. We believe that by offering exposure to labor market activity during high school, STW programs provide students with information to smooth these transitions. Moreover, by providing training for non-college bound students, these programs might reduce high school dropout rates and increase employment. Yet, if these programs supplant college preparatory courses, the marginal student could be deterred from college, or might be unprepared at enrollment.

Daniel Kreisman

Previous research on this topic is sparse due to data limitations and empirical challenges. We overcome these obstacles and build on previous work in four regards. First, we devise a formal model of STW participation. Second, we employ two novel identification strategies exploiting variation across programs and variation across states in graduation requirements to deal with selection on schools and selective participation by students. Third, we test our hypotheses using the NLSY97, including restricted data from the School Administrative Questionnaire (we have been approved to use these data by the BLS) which is uniquely suited to this research, containing detailed information on STW participation at the student and school level along with education and employment outcomes. Lastly, our estimation strategy and data allow us to compare effects across STW program types to determine which programs are most effective across student and school characteristics.