From the beginning of the “War on Poverty” in 1964, Americans have been concerned about the consequences of poverty and how to reduce its prevalence. Much public policy debate is concerned with people and families who remain poor for long periods of time. On the other hand, people who are poor for even a short period of time often require outside assistance, a “safety net,” to support themselves and their families through unfortunate circumstances.
This report uses data from the Survey of Income and Program Participation (SIPP) to examine the incidence of poverty over various time frames, from October 1991 through April 1994.1 The SIPP allows us to examine both the static and dynamic aspects of poverty. Unlike the poverty and income data collected in the Current Population Survey (CPS), which do not allow analysis of changes in income and poverty status for particular individuals, SIPP’s longitudinal data make it possible to measure movement into and out of poverty for the same people over the life of the panel. Examining poverty with longitudinal data also allows one to distinguish between short- and long-term poverty.
This report focuses on monthly measures of poverty as derived from the SIPP—namely, the percent of people who were poor in an average month of 1992 or 1993, and the percent of people who were poor 2 months or more during 1992 or 1993.2 In addition, the report examines the “chronically poor,” that is, the percent of people who were poor each month of 1992 and 1993. Median durations of the length of poverty spells for various subgroups of the population are also estimated. Finally, the report examines transitions into and out of poverty during 1992 and 1993 by estimating the number and percent of people who were poor in 1 year but not the other.3
1 The longitudinal estimates presented here are based on people who either were interviewed in all waves of the reference period, or for whom imputed information exists. Efforts were made during the life of the panel to ensure that the sample remained representative of the noninstitutional population of the United States. People who moved were followed to their new address. A “missing wave imputation” procedure was used for people who missed an interview but had completed interviews before and after the missing wave. If the people included in the estimates have different experiences of poverty than do the people who did not respond initially, left the sample, or missed two or more consecutive waves, these longitudinal estimates may be biased.
2 The SIPP collects income information and family composition on a monthly basis. Thus, poverty status can also be determined on a monthly basis by comparing monthly family income to monthly thresholds. Monthly thresholds rise from month to month by the same percentage as the Consumer Price Index (CPI). People living in families are poor if the family’s income is less than the threshold for that size family in the month. An unrelated individual is considered poor, if income in the month is less than an individual’s threshold income for the month. The number of poor in an average month is determined by averaging the number of poor in each month over the months of interest. The poverty rate for an average month is determined by summing the number of poor in each month (e.g., for calendar year 1993) and dividing by the sum of the poverty universes in each month (e.g., for calendar year 1993).
3 The transition estimates reported here are based on annual poverty estimates. Annual poverty estimates in this report are obtained by summing monthly family income over the year and comparing it to the sum of the 12 corresponding monthly thresholds, based on family size and composition in each month. These annual poverty estimates differ from the official poverty estimates that are based on the March Supplement to the Current Population Survey (CPS). The official estimates assume that the family composition, reported in March of the interview year, did not change during the previous year. Additionally, the income and poverty thresholds are updated on a yearly basis, rather than a monthly basis.
4 The figures in parentheses signify the 90-percent confidence intervals of the estimates.