PUBLISHED OR FORTHCOMING WORK: 
ECONOMICS PUBLICATIONS: 

EAT (AND DRINK) BETTER TONIGHT: FOOD STAMP BENEFIT TIMING AND DRUNK DRIVING FATALITIES (forthcoming American Journal of Health Economics)
This paper examines the relationship between the timing of food stamp benefits and daily alcohol related fatal accidents. We exploit substantial exogenous variation in state food stamp distribution dates and enrollment numbers to estimate the relationship using binary outcome and count data frameworks. Our main result is that, in contrast to previous work on income receipt and mortality, alcohol related accidents with fatalities are substantially lower on the date of food stamp receipt and that the result is largely driven by a same-day effect. Further this effect is only present on weekdays. We find no effect of receipt on non-alcohol related accidents. We hypothesize that this is possibly driven by families being more likely to eat at home on distribution days.

DOES THE TIMING OF FOOD STAMPS MATTER? A PANEL DATA ANALYSIS OF MONTHLY PURCHASING PATTERNS OF US HOUSEHOLDS (with John Gordanier, Chad Cotti and Elena Castellari  forthcoming in  Health  Economics

In this paper we examine the relationship between the timing of food stamp receipt and consumption patterns. We combine data on state distribution dates with scanner data on a panel of households. Consistent with previous work we find that purchases of a variety of goods are higher on receipt days. Additionally, we find that when receipt days are more likely to be on weekends, total monthly consumption within the same households is affected. In particular, monthly purchases of beer are higher when food stamps are distributed on a Saturday or Sunday than in months that benefits are distributed on weekdays in food stamp eligible households. For these households, total beer purchases are between 4 and 7 percent higher in those months.

DEVELOPING A SYSTEMATIC ASSESSMENT OF SCHOOL FOOD CHOICE ARCHITECTURE REVEALS KEY POINTS FOR INTERVENTION (with Melayne McInnes, Christine Blake, Edward Frongillo, and Sonya Jones) (2016 Ecology of Food and Nutrition)  
We sought to develop a structured observational method for the systematic assessment of the food-choice architecture that can be used to identify targets for behavioral economic intervention to improve children's diets. We used an ethnographic approach with observations at 12 elementary schools. Elements of the structured observational method included; decision environment; salience; accessibility/convenience; defaults/verbal prompts; number of choices; serving ware/method/packaging; and social/physical eating environment. Analysis compared observer field notes to the structured observational method. The final version of the structured observational method is divided into six sections arranged in the natural order encountered by observers. Important nudgeable components of the elementary school food-choice architecture were present, including pre-commitment, default options on the lunch line, and teacher influence in multiple areas. Our assessment method captures important structural, behavioral, and procedural
aspects of school-lunch programs that are likely to impact children's food choices. Download

IMPROVED TARGETING OF SOCIAL PROGRAMS (with Melayne McInnes, Suzanne McDermott, and Joshua Mann) (2016 Eastern Economic Journal)
In a climate of ‡at or shrinking budgets, can programs reallocate existing resources to improve effciency? We illustrate the potential for gains from redirecting resources using data from a state job coaching program that is designed to increase employment among adults with intellectual disabilities (ID). We model selection
into the program and employment outcomes for participants and non-participants allowing for potentially heterogeneous response among observationally equivalent individuals. In our simulations, we …find that state ID population employment can be increased from 10.7 percent to an upper bound of 16.7 percent by a program administrator who can allocate the job coaches to those with the most to gain. This is a 56 percent increase in the overall employment rate. While we assume that program administrators know more about individual program participants than we do, we can consider an administrator who has only the information available to the econometrician. In this case, targeting gains based only on observable characteristics would lead to 11.8 percent employment, which is an 11 percent increase in the overall employment rate. Surprisingly, a simple rule that only requires administrators to predict employment success when treated (based on observables) will achieve almost the same results. Download


WELFARE REFORM AND CHILDREN'S SHORT RUN ATTAINMENTS-A STRUCTURAL APPROACH (with Hau Chyi & Weilong Zhang)(2015 Contemporary Economic Policy)
In this paper, we use a dynamic structural model to measure the effects of (i) single mothers' work and welfare use decisions and (ii) welfare reform initiatives on the early cognitive development of the children of the NLSY79 mothers. We use PIAT-Math scores as a measure of attainment and show that both the mothers' work and welfare use benefit children on average. Our simulation of a policy that combines a time limit with work requirement reduces the use of welfare and increases employment significantly. These changes in turn significantly increase children's cognitive attainment. This implies that the welfare reform was not only successful in achieving its stated goals, but was also beneficial to welfare children's outcomes. In another policy simulation, we show that increasing work incentives for welfare population by exempting labor income from welfare tax can be a very successful policy with some additional benefits for children's outcomes. Finally, a counterfactual with an extended maternal leave policy significantly reduces employment and has negative, though economically insignificant, impact on cognitive outcomes. Download

THE ROLE OF GENDER AND SECTOR IN PROMOTION AND PAY OVER A CAREER
(with John Addison and Si Wang) (2014 Journal of Human Capital) This paper considers the role of gender and broad sector in the promotion process and thence the impact of promotion on wages, using data from the National Longitudinal Survey of Youth (NLSY79). In its use of three stages of career development, the study builds on earlier work using the NLSY79 that tackles gender differences in early career alone. The raw data suggest findings broadly favorable to females in terms of promotion over a career. But when we examine gender differences by educational groups the advantages seem to be confined to less educated females. And while there is a strong return to education both in terms of promotion probability and attendant wage growth for males in the later career years this is not the case for educated females. While these findings are consistent with discrimination they may also reflect a choice among educated females for flexibility over career. Download

JOB PROMOTION IN MID-CAREER: GENDER, RECESSION and 'CROWDING' (with John Addison and Si Wang) (2014 Monthly Labor Review
Data from the National Longitudinal Survey of Youth 1979 indicate that between 1996 and 2010 females on average lost some of the promotion momentum they had achieved at the beginning of mid-career, although they outperformed males in this regard. For both genders economic downturn has contributed to reduced promotion probabilities. In the case of women, however, cohort effects rather than the cycle seem to explain the promotion experience during the Great Recession. Promotions translate into higher real wage increases, and typically more so where job responsibilities increase. Crowding effects, if not necessarily a thing of the past, are no longer manifested in reduced female promotion rates or earnings.

WELFARE AND WORK PARTICIPATION OF SINGLE MOTHERS AND CHILDREN’S ATTAINMENTS  (with Hau Chyi) (Economic Inquiry, 2013)
We examine the effects of single mothers' welfare use and employment decisions on children's short-run cognitive development, as measured by their preschool standardized math test scores. We control for three mechanisms through which these decisions might affect children's outcomes: direct monetary benefits, parental time invested in the child, and non-pecuniary benefits from in-kind transfer programs such as Medicaid. We employ a correction function approach and control for state fixed effects to address the endogenous nature of welfare participation and employment decisions. Our estimates suggest that although each additional quarter of either mother's employment or welfare use results in only a small increase in a child's standardized math test score, the total effects after several quarters are sizable. We allow mothers' decisions to have varying effects on attainment by children's observed innate ability and by the intensity of welfare use and employment. A child who has the mean level of observed innate ability with a mother who simultaneously worked and used welfare in all twenty quarters after childbirth experiences an 8.25-standardized-point increase in standardized scores. The positive impact is more pronounced for the more disadvantaged children, who tend to be born to mothers with low Armed Forces Qualification Test (AFQT) scores, or have lower birth weights. We also examine the effects using timing of employment and welfare use, as well as children's maturity and gender. Download


MINIMUM WAGES, MARKET INFLEXIBILITIES AND FEMALE EMPLOYMENT IN SELECT OECD COUNTRIES (with John Addison) (Industrial and Labor Relations Review, 2012)
In this paper we investigate the employment consequences of minimum wage regulation in 16 OECD countries, 1970-2008. Our treatment is motivated by Neumark and Wascher’s (2004) seminal cross-country study using panel methods to estimate minimum wage effects among teenagers and young adults. Apart from the longer time interval examined, a major departure is our focus on prime-age females, a group often neglected in the minimum wage literature. Another is our deployment of time-varying policy and institutional regressors. The average effects we report are consistent with minimum wages causing material employment losses among the target group. Indeed, higher minimum wages are also associated with elevated joblessness, although these unemployment effects are less precisely estimated. Further, although the we find common ground with Neumark and Wascher as regards the role of some individual labor market institutions and policies, we do not observe the same patterns in the institutional data. Specifically, prime-age females do not exhibit stronger employment losses in countries with the least regulated markets. Download


CONFESSIONS OF AN INTERNET MONOPOLIST: DEMAND ESTIMATION FOR A VERSIONED INFORMATION GOOD (with Henry Chappell and Paulo Guimarães) (Managerial and Decision Economics, 2011)
We develop and apply a method for estimating demand system parameters for versioned information goods. Our analysis uses data collected from a web-based field experiment in which prices and versions of an information good were exogenously varied. Using a maximum simulated likelihood (MSL) procedure, we estimate parameters characterizing distributions of utility functions over a population of potential buyers. We then construct profit-maximizing versioning and pricing plans for the seller, and assess the welfare implications of those plans. Because firms increasingly have opportunities to collect information by tracking behavior of customers, methods similar to ours are likely to prove useful in future commercial applications. Download

DOES SUPPORTED EMPLOYMENT WORK?(with Melayne McInnes, Suzanne McDermott, and Joshua Mann)(Journal of Policy Analysis and Management, 2010)
Providing employment-related services, including supported employment through job coaches, has been a priority in federal policy since the enactment of the Developmental Disabilities Assistance and Bill of Rights Act in 1984. We take advantage of a unique panel data set of all clients served by the SC Department of Disabilities and Special Needs between 1999 and 2005 to investigate whether job coaching leads to stable employment in community settings. The data contain information on individual characteristics, such as IQ and the presence of emotional and behavioral problems, that are likely to affect both employment propensity and likelihood of receiving job coaching. Our results show that unobserved individual characteristics and endogeneity strongly bias naive estimates of the effects of job coaching. However, even after correcting for these biases, an economically and statistically significant treatment effect remains. Download


DETERMINANTS OF HEALTH CARE UTILIZATION AMONG YOUNG ADULTS WITH MUSCULAR DYSTROPHY: ROLE OF SOCIOECONOMIC STATUS  (with Suzanne McDermott, Joshua Mann, James Harden, Julie Royer, Lijing Ouyang) (2014 Medical Care)
For people with Muscular Dystrophy (MD) health care access is crucial and utilization is expected to be high. A multidisciplinary approach is needed for optimal management of symptoms of this rare condition. Regular primary care, specialty care, therapy and medicine use can improve quality of care and reduce need for emergency treatment (ER) and hospitalization. We analyzed health insurance and administrative data to assess the impact of socioeconomic status (SES) on regular care for MD patients. We used South Carolina Medicaid and other administrative data for individuals 15 to 24 years of age to determine annual health care utilization patterns for individuals with MD, conditional on economic well-being (as indicated by food stamps receipt), failure to complete high school, and minority racial status. We studied adolescents and young adults with MD because this age group represents a time when the condition is typically intensifying and the transition from pediatric to adult care is expected. We provide unconditional descriptive comparisons, and we also used GEE models to analyze utilization data.Socioeconomic status (SES) is correlated with health care utilization among adolescents and young adults with MD. Low SES in general implies lower overall utilization, and less primary care, therapy and specialist care use but higher incidence of hospitalization and ER use. We see that these differences become more significant as our subjects transition to adult care from pediatric care. Even in regression analysis, where we take into account individual unobserved factors and allow clustering at the individual level, most of these differences are still present in conditional means and are in most cases statistically significant. As there are differences in health care utilization by SES even when individuals have access to the same health care benefits, simply offering coverage to individuals with MD may not be sufficient in eliminating health disparities. Differences in health care use may be resulting from disparities in resource awareness, health knowledge, and access barriers such as transportation. As ER and inpatient care is significantly more costly than primary care, therapy, and specialist care, interventions to modify utilization patterns have the potential to be cost efficient. Download


WORKING PAPERS UNDER REVIEW:

FOLLOWING (NOT QUITE IN YOUR FATHER'S FOOTSTEPS: TASK FOLLOWERS AND LABOR MARKET OUTCOMES  (with John Gordanier and Liwen Chen)

This paper examines the extent to which children enter into occupations that are different from their father’s occupation, but require similar skills, which we call task following. We also consider the possibility that fathers are able to transfer task specific human capital either through
investments or genetic endowments to their children. We show that there is indeed substantial task following, beyond occupational following and that task following is associated with a wage premium of around 5% over otherwise identical workers employed in a job with the same primary task. The wage premium is robust to controls for industry, occupation categories and occupation characteristics. The premium is largest for followers in non-routine cognitive jobs and college graduates.

WHEN DOES IT COUNT? THE TIMING OF FOOD STAMPS RECEIPT AND EDUCATIONAL PERFORMANCE IN MATHEMATICS (with John Gordanier and Chad Cotti)

This paper exploits administrative data that allows us to match food stamp benefit receipt dates to student math test scores. Our main results are that scores are slightly worse when the length of time since benefit receipt changes marginally. However, they are notably lower when the exam falls much later in the benefit cycle. Further, when food stamps arrive on the weekend immediately preceding the exam, estimates show that scores are significantly lower. We also find evidence of negative spillovers from weekend receipt on performance. Our results provide evidence that households are unable to sufficiently smooth consumption and that this has measurable effects on student performance. The fact that weekend receipt differs suggests a behavioral response from households that also has meaningful effects.

MISSION STATEMENTS, LOCATION CHOICE AND THE SUPPLY OF CHARTER SCHOOLS (with Crystal Zhan)

Charter schools represent one of the most popular yet contentious education policy issues in the United States today. In this paper, we link the mission statements of charter schools to the characteristics of traditional public schools, students, and neighborhoods in the surrounding areas to better understand the supply of charter schools. We examine the charter schools in six states: Arizona, Georgia, New York, Ohio, Oregon, and Pennsylvania. Empirical analysis suggests that (1) charter schools that emphasize academics, in particular, core subjects and college preparation, prefer locations with low-performing traditional public schools and more underserved student population; (2) charter schools that offer bilingual programs are more likely to emerge in ethnically diverse neighborhoods. These results may verify that the supply decisions of charter schools in terms of types of educational programs respond to the specific demands of students and parents. The paper may also shed light on the parental preferences toward schooling options conditional on existing public schools and their demographic backgrounds.

WORKING PAPERS : 

RISK TOLERANCE, GENDER AND PROMOTION (with Melayne McInnes and Si Wang) 
Field and experimental studies find that women are more risk averse then men, which begs an important empirical question: how much do risk attitude differences by gender explain labor market outcome differences? We investigate one aspect of the labor market, mid-career promotion, which has been shown to be a critical component of the wage determination process. We use the core cohort of the NLSY79 which provides a nationally representative panel of data for the cohort of individuals aged 14 to 22 years in 1979. We find that being moderately or highly tolerant to risk increases one's promotability for both males and females; however, these effects disappear once we control for unobserved heterogeneity for females. For males we see a significantly higher promotion probability for being moderately risk tolerant even when we control for individual unobserved heterogeneity and industry and occupational characteristics.

THE OCCUPATIONAL FEMINIZATION OF PROMOTIONS (with John Addison, Dan Qu and Si Wang) 
Are women more or less likely to be promoted in environments where they are in gender minority? Evidence from existing studies differ by occupation and by the way promotion is defined. In this project, we construct and use a nationally representative data set, not only a sample of a single industry or occupation. We find while women are slightly more likely to be promoted than men, the occupational feminization does not affect one’s promotion probability. Women in “female jobs”, however, are less likely to be promoted than their male co-workers. Even with occupational controls these effects remain. 


RARE CONDITIONS PROJECT PAPERS : 

EMERGENCY DEPARTMENT VISITS AND INPATIENT HOSPITALIZATIONS FOR ADOLESCENTS AND YOUNG ADULTS WITH FRAGILE X SYNDROME  (with Suzanne McDermott, Joshua Mann, James Harden, Julie Royer, Lijing Ouyang) (forthcoming  in American Journal of Intellectual and Developmental Disabilities

CO-MORBID CONDITIONS ASSOCIATED WITH ADOLESCENTS AND YOUNG ADULTS WITH FRAGILE X SYNDROME  (with Xin Tong, Suzanne McDermott, Joshua Mann, Julie Royers, James Hardin and Lijing Ouyang) (forthcoming in Journal of Autism and Developmental Disorders

USING STATE ADMINISTRATIVE DATA SOURCES TO STUDY ADOLESCENT AND YOUNG ADULTS WITH RARE CONDITIONS (with Suzanne McDermott, Joshua Mann, Julie Royers, James Hardin, Lijing Ouyang and Julie Bolen)(forthcoming in Journal of General Internal Medicine)


CO-MORBID CONDITIONS ASSOCIATED WITH ADOLESCENTS AND YOUNG ADULTS WITH FRAGILE-X SYNDROME (with Xin Tong, Suzanne McDermott, Joshua Mann, Julie Royers, James Hardin and Lijing Ouyang)  (Focus on Autism and Other Developmental Disorders)

HOSPITAL VISITS IN YOUNG PEOPLE WITH SPINA BIFIDA  (with Joshua R. Mann, Julie A. Royer, Margaret A. Turk, Suzanne McDermott, Margaret M. Holland, James W. Hardin, Judy K. Thibadeau) (American Journal of Physical Medicine & Rehabilitation)

SKIN ULCERS AND MORTALITY AMONG ADOLESCENTS AND YOUNG ADULTS WITH SPINA BIFIDA IN SOUTH CAROLINA DURING 2000-2010 (With Bo Chai, Suzanne McDermott, Joshua Mann, Julie Royer, James Hardin, Yinding Wang and Lijing Ouyang) Revise and Resubmit Journal of Child Neurology


IDEAS:

OBESITY AND FOOD STAMPS (with T.J. Classen and Hau Chyi)
We use previous estimates of responses to food insecurity and consider whether increases in food budgets from the food stamp program (FSP) result in substitution toward healthier foods, possibly due to income effects of benefits. This allows us to consider policy experiments in which constraints on possible food choices available from the FSP are imposed to require certain minimal nutritional requirements of covered foods. To estimate substitution effects of policy changes that limit the set of possible food choices based on nutritional content and other potential changes to the FSP, we employ the USDA's Continuing Survey of Food Intake by Individuals (CSFII). These are detailed food diary data that have previously been used to consider the nutritional choices of FSP participants (e.g., USDA study by Gleason et. al. (2000)). The policy exercises include targeting the benefit to a smaller set of food choices, perhaps based on nutrient content or other nutritional measures (%of calories from fat, etc.). The food security issue and substantial increase in consumption around the benefit might be increased if the benefit were limited to increase more fresh foods which are not generally storable. Another policy experiment considers the effect of offering benefits on a more frequent basis. This is based on the food insecurity issues and cyclical consumption patterns identified in Shapiro (2005). 

MARRIAGE BEHAVIOR IN RURAL AND URBAN CHINA (with Hau Chyi and Wenjian Zhao) 
This paper investigates the differences in the marriage markets in rural and urban China. Using data on Chinese men and women from the CHNS, important roles for asset level and current income in the marriage transition are found. There are significant differences between rural and urban marriage markets arising from the restrictions on rural-urban migration and gender imbalance of the population in China. For rural and urban women, high current income increases the probability of marriage while little impact is found for asset level. Both high asset level and current income increase the probability of marriage for rural men, and they have no significant effects on marriage transition for urban men. 

EMPLOYMENT EFFECTS OF MINIMUM WAGES IN MARKETS WITH FIXED EMPLOYMENT COSTS 
This paper structurally models and estimates the employment effects of a minimum wage regulation in an inflexible labor market with fixed employment costs. When there are …fixed costs associated with employment, minimum-wage regulation not only results in a reduction in employment among low-productivity workers but also shifts the distribution of hours for the available jobs in the market, resulting in a scarcity of part-time jobs. Thus, for sufficiently high employment costs, a minimum wage makes it less likely for "marginal" workers to enter and stay in the labor market. I estimate the model using survey data from Turkey. I find a significant reduction in employment due to the loss of part-time jobs caused by the national minimum-wage policy in this highly inflexible labor market. Download

IDENTIFYING LOVELY JOBS: JOB QUALITY AND ECONOMIC INDEPENDENCE OF WELFARE USERS (with Hau Chyi)
Recent work in the welfare literature suggests that a repetitive, low-paying job may not sustain the long-term welfare independence of low-skilled single mothers. In this study, we use low-skilled women from the SIPP 1990, 1991, 1992, 1993, and 1996 panels to study the .effect of occupation choice on the probability of leaving welfare. Jobs are grouped by their employment and earning potential characterized by the Dictionary of Occupational Titles. We and that, conditional on working, single mothers who work in a job with higher employment and earning potential (the so-called good jobs) are more likely to be welfare independent than those who are employed in bad jobs. Download


OIL PRICE HIKE, RURAL DRIVING AND EMPLOYMENT ARRANGEMENTS 
Within last year crude oil prices doubled and the upward trend of gasoline prices has been a talking point for more than a year. Combined with economic downturn of the last couple months this price increase pattern, we propose, has interesting implications for rural job markets where people usually drive long distances to commute between work and home. These workers are also low productivity workers mostly working at about minimum wage level jobs. Moreover, they are more likely to have low m.p.g. cars. When this is the case, increase in cost of driving constitutes an increase in the "fixed cost" of employment for a worker, which makes it less attractive to work short hours (part time) or to work at all. In general part-time jobs are countercyclical, that is their prevalence is higher during economic downturns. Which, I propose, may not be the case at this time and may have implications on the general time use of the workers and the organization of the workplace and work hours.

HOUSING CRISES AND WORKERS' IMMOBILITY: LONG RUN INDIRECT CAREER EFFECTS OF THE DOWNTURN IN THE HOUSING MARKET
"The rapid decline in housing prices is distorting the normal workings of the American labor market. Mobility opens up job opportunities, allowing workers to go where they are most needed. When housing is not an obstacle, more than five million men and women, nearly 4 percent of the nation's work force, move annually from one place to another --- to a new job after a layoff, or to higher-paying work, or to the next rung in a career, often the goal of a corporate transfer. Now that mobility is increasingly restricted many workers may be missing the opportunities to make the early career moves which literature shows has great impact on long run wage growth. Unable to sell their homes easily and move on, tens of thousands of people are making the labor force less flexible just as a weakening economy puts pressure on workers to move to wherever companies are still hiring. No government agency counts those who move for a job, either across state lines or just from one town to another in the same state. The Census Bureau, however, calculates how many people move across state lines for all reasons, and that number fell by a startling 27 percent last year, after climbing by almost that percentage for each of the previous three years. With homes changing hands easily in a booming market, interstate migration reached 2.2 million people in 2006, excluding the effects of Hurricane Katrina. As the economy and home prices began to unravel in 2007, however, interstate migration plunged to 1.6 million people."(NY TIMES)
To be able to disentangle the impact the general economic downturn has on the labor market from the effect of the housing crises a detailed data of housing prices is needed, in addition to income, labor market and expenditure information. This project will involve a dynamic modeling of the labor market moves of individuals in relationship to migration and housing consumption decisions, which requires a long time series, which also is provided with CES data.

FOOD PRICES AND OBESITY(with Tim Classen)
Drewnowski (2000) compares the prices of 370 foods sold at supermarkets in the Seattle area and shows that per calorie, junk foods not only cost less than fruits and vegetables, but junk food prices also are less likely to fluctuate. This can explain to some extent why the highest rates of obesity are seen among people in lower-income groups. Price differences across food groups are substantial: Energy-dense low nutrition food costs on average $1.76 per 1,000 calories, compared with $18.16 per 1,000 calories for low-energy but nutritious foods. The survey also showed that low-calorie foods were more likely to increase in price, surging 19.5 percent over the two-year study period. High-calorie foods remained a relative bargain, dropping in price by 1.8 percent. Based on his findings, a 2,000-calorie diet would cost just $3.52 a day if it consisted of junk food, compared with $36.32 a day for a diet of low-energy dense foods.

These facts are important for the research on obesity. One would expect income to be positively correlated with the prevalence of obesity since food being a normal good, we expect consumption of it to increase with income. However, looking at the data it is clear that obesity is a huge epidemic in USA mostly among low income families. This observation together with the data suggests low income people may be consuming more of the high calorie low price foods as alternatives to healthier low calorie items.

MATERNAL EMPLOYMENT AND BMI OUTCOME OF CHILDREN (with Hau Chyi and Wei Lin) 
As it is well documented in literature that parents’ socioeconomic status is highly correlated with their child's health. In this paper we study the impact of maternal employment on child weight outcomes by exploiting the dramatic decline of labor force participation rate in China since 1997, when many state-owned enterprises laid off many workers. We use the matched household-child data from the China Health and Nutrition Survey (CHNS) and employ several econometric models, including discrete choice models with household fixed effect, and an IV, to handle the endogeneity of mother’s work choices. Our preliminary results indicate that low-income families are more likely to have underweight children after controlling the maternal employment; and families with non-working mother are more likely to have to have obese children after controlling the household income. Moreover, rural families with grandparents in the household are inclined to have obese children, while this relationship does not show up in urban families.