A method for estimating child poverty rates, projections for the short-term and the relationship between child poverty and child care subsidy receipt at the county level

Abstract

At present, county level annual poverty rates for broad age groups at different poverty thresholds are not readily available. Unlike the 2000 census, the 2010 census did not collect income data to allow calculation of poverty rates, and future decennial censuses will not. Thus, there is a need for a method to estimate annual poverty rates at the county level. To fill in the existing gap, this paper develops a method conceptually similar to the regression-based modeling approach utilized for Census Bureau Small Area Estimates (SAIPE). This is applied to estimate and project the proportion of children less than 12 years of age whose income is below the 200% poverty threshold, by county in New York State from 2008 to 2010. Multiple regression techniques including random effects and Auto Regressive Integrated Moving Average (ARIMA) models are applied. Data obtained from the U.S. Census Bureau, the American Community Survey and administrative sources are utilized. Administrative data sources adjust for, incorporate and model recent changes in the county economic situation. Finally, this paper contributes to the literature on county level relationships between poverty rates and income-based benefits, particularly on child care subsidies. Specifically, the association between the proportion of children receiving child care subsidies and child poverty rates vary by metropolitan, micropolitan and non-core-based designation of counties.