Can We Fix the Federal Poverty Measure so It Provides Reliable Information about Changes in Children's Living Conditions?

The official poverty rate for American children rose during the 1970s and 1980s and fell during the 1990s, but it was still significantly higher in 1999 than in 1969. Since the poverty line is supposed to represent a fixed standard of living, these change in child poverty imply that low-income children’s standard of living was lower in 1999 than in 1969. We test this hypothesis using data on housing conditions, telephone service, automobile ownership, doctor visits, and food consumption. With the possible exception of food consumption, trends in these domains do not match the trend in official child poverty. Plausible corrections to the way poverty is measured generate a time series that comes closer to matching trends in living conditions, but the match is still far from perfect. CAN WE FIX THE FEDERAL POVERTY MEASURE SO IT PROVIDES RELIABLE INFORMATION ABOUT CHANGES IN CHILDREN'S LIVING CONDITIONS? The federal government established its first official poverty measure in 1969. This measure, based on the work of Mollie Orshansky (1965), defined a family as poor if its money income in any given calendar year fell below a threshold that varied with the family's size and composition. Since 1969 the poverty threshold has risen at the same rate as the Consumer Price Index for urban consumers (CPI-U) but has not changed much in other respects. In principle, therefore, families with incomes at the poverty line are meant to have the same standard of living today that they had in 1969. According to the Census Bureau’s “official” poverty series, the poverty rate for all persons hardly changed between 1969 and 1999. The official rate was about 12 percent in both 1969 and 1979, 13 percent in 1989, and 12 percent in 1999 (U.S. Census Bureau, 2000:B-12). The poverty rate for children has risen. It was 14 percent in 1969, 16 percent in 1979, 20 percent in 1989, and 17 percent in 1999. These statistics imply that low-income children were worse off in 1999 than 1969. This paper investigates whether that was in fact the case. The stability of the overall poverty rate reflects increases for some groups (notably children) and declines for other groups (notably the elderly), so a convincing assessment of the match between trends in poverty and trends in living conditions would require us to compare these groups. This paper, which focuses on child poverty, is a first step in that direction. We begin with children partly because the increase in child poverty during the 1970s and 1980s was widely cited as evidence of policy failure. In addition, the large increase in child poverty during the 1970s and 1980s followed by a reversal of this trend during the 1990s makes it relatively easy to see whether direct measures of living conditions follow the same trajectory. We should emphasize at the outset that we will not discuss whether the official poverty line ought to represent a fixed standard of living or falling rather than rising in tandem with the overall standard of living of the larger population. Both approaches provide useful information, and we see no compelling reason to choose one to the exclusion of the other. Here, however, we assume that the poverty line is supposed to