The Dynamics of Poverty Among the Elderly :

This paper examines the risk of poverty among a sample of couples and widows interviewed in the Retirement History Survey over a ten-year period. The longi tudina1 na ture of the da ta allows us to look a t the dynamics of poverty among older married women as they age and become widowed. We arrange the data to simulate a cross-section and then compare this method of calculating poverty rates with a longitudinal analysis of the actual pattern of poverty among the households in our sample. Cross-sectional data can only identify net changes in poverty--whether a household is or is not poor in any particular year. When individual movements into and out of poverty over a period of time are identified, the risk of becoming poor during that time period is more than double the highest annual risk among elderly persons who stay married, and is raised by 25 percent for widows. Both couples and widows move into and out of poverty over the tenyear period, indicating that our stereotype of the static nature of poverty among the elderly must be altered. We also calculate exit and reentry rates for poor elderly households. Surprisingly, those rates among couples are not significantly different from the rates among widows. Over 80 percent of both widows and couples exit poverty after three (two-year) periods, and just over one-third reenter after the same number of periods. We also look at the relationship between widowhood and poverty. We find that when their husbands were alive, those women who subsequently became widows were more likely to be poor than were couples in which the husband remained alive throughout the survey period. Hence the importance of widowhood as an explanation of poverty is somewhat overstated by simply comparing married and widowed groups. The Dynamics of Poverty Among the Elderly: Income Transitions at Older Stages of Life A major accomplishment of federal policy over the last two decades has been to increase the well-being of the aged. Despite this overall improvement, poverty continues to afflict some elderly persons. That the fastest-growing population of poor households consists of those headed by women is well documented (Bane, 1984). Older widows make up a large subgroup in this category. According to the Current Population Survey, in 1980 over 32 percent of all women aged 65 and older who were living alone were poor, a poverty rate more than twice that for the aged in general (U.S. Bureau of the Census, 1982). And widowed, divorced and single women account for over 50 percent of all aged poor.' Though poverty rates have fallen for older women over recent decades, women are less likely to have their incomes raised by economic growth or by the public and private income transfer programs that have moved couples out of poverty (Ross, 1984). The risk of poverty among couples and single men has sharply fallen, leaving poverty in old age a characteristic primarily of older women. Despite the gr~ater risk of being poor that is faced by older widows, we know surprisingly little about the timing, duration and causes of poverty among this group. Single-year, cross-sectional income data provide information on the percentage of the population that is in the midst of a spell of poverty at that moment, but can provide only a crude approximation of the total risk of poverty as people age. Such data provide no information on the timing or the length of a poverty spell beyond what can be inferred from the poverty rates that prevail among different age groups, a crude method that assumes constant age-related poverty changes across cohorts.