Understanding the link between poverty and food insecurity among children: Does the definition of poverty matter?

Abstract

This paper examines the association between poverty and food insecurity among children, using two different definitions of poverty – the official poverty measure (OPM) and the new supplemental poverty measure (SPM) of the Census Bureau, which is based on a more inclusive definition of family resources and needs. Our analysis is based on data from the 2001–2011 Current Population Survey and shows that food insecurity and very low food security among children decline as income-to-needs ratio increases. The point estimates show that the associations are stronger as measured by the new supplemental measure of income-to-needs ratio than when estimated through the official measure. Statistical tests reject the hypothesis that poor households' odds of experiencing low food security are the same whether the SPM or OPM measure is used; but the tests do not reject the hypothesis when very low food security is the outcome.

Keywords:

• food insecurity among children
• poverty
• official poverty measure
• supplemental poverty measure

Acknowledgments

The authors thank James Ziliak and Craig Gundersen for their thoughtful comments on an earlier draft and also thank Liana Fox, Nathan Hutto, and Chris Wimer with their help with the SPM thresholds.

Funding

This project was supported with a grant from the University of Kentucky Center for Poverty Research through funding by the US Department of Agriculture, Food and Nutrition Service, contract number [AG-3198-B-10-0028]. The opinions and conclusions expressed herein are solely those of the author(s) and should not be construed as representing the opinions or policies of the UKCPR or any agency of the Federal Government. We are also grateful for support from Grant R24, Project Number [HD058486-03] from the National Institute of Child Health and Human Development to the Columbia Population Research Center.

Notes on Contributors

Vanessa Wight is an associate research scientist at the Columbia University School of Social Work. She is a sociologist and her research interests include family demography, poverty, and child well-being. She is also an affiliate with the Columbia University Population Research Center.Neeraj Kaushal is an associate professor at the Columbia University School of Social Work, and research associate at the National Bureau of Economic Research. She is also a research fellow at IZA - the Institute for the Study of Labor, Bonn, Germany. She is a labor and health economist, and her research focus is on how policies and events affect the well-being of low-income families with special emphasis on immigrants.Jane Waldfogel is the Compton Foundation Centennial Professor for the Prevention of Children's and Youth Problems at the Columbia University School of Social Work and a visiting professor at the Centre for Analysis of Social Exclusion at the London School of Economics. She has written extensively on the impact of public policies on child and family well-being; her current research includes studies of work–family policies, improving the measurement of poverty, and understanding social mobility across countries.Irwin Garfinkel is the Mitchell I. Ginsberg Professor of Contemporary Urban Problems at the Columbia University School of Social Work and co-founding director of the Columbia Population Research Center. A social worker and an economist by training, he has authored or co-authored over 200 scientific articles and 16 books and edited volumes on poverty, income transfers, program evaluation, single parent families and child support, and the welfare state.

Notes

1. Research suggests that maternal depression, mental health disorders, and intimate partner violence are associated with food insecurity, but the direction of causality is not clearly established (Casey et al. 2004; Corcoran, Heflin, and Siefert 1999; Ettner 1996; Lohman et al. 2009; Whitaker et al. 2006a; Sareen et al. 2011). It is possible that food insecurity increases the risk of domestic violence, maternal depression, and mental health disorders; at the same time, this latter set of factors may increase the risk of experiencing food insecurity.

2. Low income is almost inherent in the definition and measurement of food insecurity. Indeed, in the 18-item Household Food Security Scale, many of the questions ask whether a given behavior or condition occurred because of lack of money.

3. Assets, particularly liquid assets, are protective against food insecurity (Leete and Bania 2010; Ribar and Hamrick 2003). Nord and Brent (2002) find uneven income, changes in household composition, multiple families residing in the same household without fully sharing resources, and unexpected economic needs (e.g., unexpected medical expenses) to be possible explanations for the anomalies in the relationship between income and food insecurity.

4. Originally based on data from the 1950s, the poverty threshold was set at three times the cost of food and adjusted for family size. Since then, the measure has been updated only for inflation. Yet food now comprises only about one-seventh of an average family's expenses, while the costs of housing, child care, health care, and transportation have grown disproportionately.

5. Research has shown that the food-security scale is stable over time and across broad subgroups of the population and is externally valid (Hamilton et al. 1997; Ohls, Radbill, and Schirm 2001).

6. We also imputed a measure of continuous income from the March ASEC to the December CPS using a regression-based method that estimates continuous income separately by year and family income band in the March CPS. The results from the regression-based method, which are available upon request, are similar to what we report in this paper, suggesting that the relationship between income-to-needs and food insecurity among children is very similar from the two specifications of income – median income and imputed income. We have elected to present results from the former.

7. The CPS uses a rotating sample in which respondents are interviewed for four consecutive months, sit out for 8 months, and are then interviewed again for 4 months before being permanently retired. Thus, for about one-fourth of our December sample, data drawn from the March CPS and used in constructing the SPM measure were taken from the same households that comprise the sample analyzed in this paper.

8. In addition, the SPM sets separate thresholds for families who own their homes free and clear and families who either have mortgages or rent, and it adjusts thresholds for family size and composition and geographic differences in housing costs. In this analysis, we adjust thresholds for family size and composition. However, due to data limitations, we do not set separate thresholds for families who own their homes free and clear and families who have mortgages or rent, nor do we adjust for geographic differences in housing costs.

9. This is a simplification of complex United States Department of Housing and Urban Development (HUD) guidelines, but modeling the more complex HUD guidelines would require more information than is available in the March CPS. We apply a small correction factor to match our estimates of housing-subsidy values with the Census estimates.

10. In results not reported, we also estimated an ordered logistic regression model. The results are similar to what is reported below.

11. We tested whether coefficients across models were statistically significantly different by dividing the difference between the coefficients by the square root of the sum of the sampling variances of the two coefficients.

12. Betson (1996) argues that the equivalence scales that are used to adjust the poverty thresholds for families of different size and composition are inadequate as they do not either fully take into consideration economies of scale or adjust for differences in consumption patterns. Thus, in results not shown, very low food security among children was regressed on family income in deciles. The results from this logistic regression are consistent with what is reported in the paper. Income is strongly associated with very low food security among children. Controlling for a rich set of covariates including number of young children, number of children aged 6–18, number of adults, and number of elderly, children in the bottom income decile are 12 times as likely as children in the top income decile to experience very low food security. The odds of experiencing very low food security decline as income increases.

Additional information

Funding

Funding: This project was supported with a grant from the University of Kentucky Center for Poverty Research through funding by the US Department of Agriculture, Food and Nutrition Service, contract number [AG-3198-B-10-0028]. The opinions and conclusions expressed herein are solely those of the author(s) and should not be construed as representing the opinions or policies of the UKCPR or any agency of the Federal Government. We are also grateful for support from Grant R24, Project Number [HD058486-03] from the National Institute of Child Health and Human Development to the Columbia Population Research Center.