We use postcensal estimates of the resident population as predictor variables in the county models of the number of people in poverty. The use of both population and the number of people represented on tax returns as predictor variables in these models provides an implicit measure of the number of persons omitted by tax returns, many of whom may have low incomes. In the state-level models, the complement of the ratio of the number of people represented on tax returns to the resident population plays the same role.
In the state-level models, the dependent variable, and the variable predicted for each state, is the ratio of numbers of people in poverty to population, as measured in the Annual Social and Economic Supplement (ASEC) of the Current Population Survey (CPS). To transform these into estimated numbers of people in poverty, we multiply each estimated ratio by a demographic estimate of the population as covered by the ASEC. The ASEC universe includes the civilian noninstitutional population of the United States and members of the armed forces in the United States living off post or with their families on post. It excludes all other members of the armed forces and treats college students as residing in their parental homes. To adjust estimates of total resident population to the ASEC universe, we subtract unpublished demographic estimates of the group quarters population by age and the appropriate type of group quarters from the estimate of the total resident population. Prior to income year 1999, we also added an estimate of "net student migrants" to the appropriate age group. We no longer make that adjustment because it is no longer a part of the demographic population estimates. Finally, we use estimates of the poverty universe at both the state and county level to compute the percentages of persons in poverty shown in the tables of estimates. We form poverty universe estimates from estimates of the resident non-institutional population by adjusting them to exclude several other population subgroups (e.g., foster children under age 15) and to limit the estimates of the number of children to related children. We describe these adjustments in more detail in the section on Denominators for State and County Poverty Rates: 1993-2004.
More on the Population Estimates Program.