Into the Workforce and Out of the Home? A Study of Maternal Employment and Food Purchased for At-Home Consumption

A robust body of literature suggests that maternal employment is positively correlated with childhood obesity (Anderson et al., 2003; Ruhm, 2008; von Hinke Kessler Scholder, 2008; Anderson, 2011; Bishop, 2011; Greve, 2011). Studies have examined parental supervision, physical activity, and meals away from home as potential mechanisms for this correlation. On average, children with working mothers watch more television and consume more unhealthy foods (Chia, 2008; Datar et al., 2014). Additionally, working mothers spend fewer minutes per day grocery shopping, preparing meals, and eating with children (Cawley and Liu, 2012). Although these findings are consistent with the observed correlation, time use does not fully explain the relationship between mother's work status and children’s body mass index (Morrissey et al., 2011; Gwozdz et al., 2013; Ziol-Guest et al., 2013).

I extend this body of literature by looking for other mechanisms for the relationship between maternal employment and childhood obesity within food purchased for at-home consumption. Because less than half of total food expenditures are for meals away from home (USDA, 2012), food purchased for at-home consumption may also significantly contribute to childhood obesity. However, the effect of maternal employment on food purchased for at-home consumption is theoretically ambiguous. Households may substitute meals away from home or increase purchases of energy-dense prepared foods due to time constraints (substitution effect), or they may increase purchases of higher quality nutritious foods with increased income (income effect).

I exploit exogenous variation in maternal employment associated with the youngest child’s eligibility for kindergarten to identify the effect of maternal employment on food purchased for at-home consumption (Gelbach, 2002; Morrill, 2011). This variation is the result of the youngest child’s date of birth relative to the state-year specific kindergarten eligibility cutoff date. The identifying assumption is that the youngest child’s eligibility for kindergarten is uncorrelated with household food purchases. To satisfy this assumption, I restrict my sample to households with children ages 2 to 8 years old and I estimate the effect of maternal employment on average grocery expenditures during summer months when children are not enrolled in school.

I use Nielsen Consumer Panel (NCP) data for years 2004-2013 matched to state-year specific kindergarten eligibility cutoff dates to estimate the impact of maternal employment on food purchased for at-home consumption, overall and by product category. The NCP data include approximately 40,000-60,000 nationally representative households annually that report UPC-level information on retail purchases. Households participating in the NCP complete a demographic survey in the fall (October-December) preceding the year they report their retail purchases.

My preliminary results are consistent with prior research on maternal employment and suggest that substitution effects dominate income effects. On average, working mothers purchase more energy-dense prepared foods, desserts and snacks, and sugar sweetened beverages. Further, households with working mothers purchase fewer fruits and vegetables, and they purchase less juice and dairy. The results from this study identify a new mechanism by which maternal employment may affect childhood obesity and they further highlight the impact of time constraints on working families.