Do Official Poverty Rates Provide Useful Information about Trends in Children's Economic Welfare?

Comparing business cycle peaks, the official poverty rate for American children rose from 14.0% in 1969 to 19.6% in 1989. This increase has been widely cited in policy debates, both as evidence that the war on poverty was counterproductive and as evidence that it should be intensified. But this trend estimate is upwardly biased for at least three major reasons: 1) children's households include more nonrelatives, whose income is not counted when the Census Bureau decides whether a child is poor; 2) official price indices overstate inflation; and 3) Medicaid, food stamps, and rent subsidies have raised many children's material standard of living without raising their household income. In addition, low-income households probably have somewhat more unreported income today than in the past. When we correct the first three problems, child poverty appears to fall sharply during the 1970s and remain roughly constant during the 1980s. Measures of "consumption poverty" among children also increase less than measures of "income poverty" over the same interval. Direct measures of children's living standards mostly support this "revisionist" account. Low-income children saw doctors more often, lived in less-crowded housing, and were more likely to have telephones in 1980 than in 1970. These indicators did not change much between 1980 and 1990. Low-income children's homes were also more likely to have complete plumbing, electrical outlets in every room, a modern sewage system, and air conditioning in 1990 than in the early 1970s. Home ownership declined among low-income parents during both the 1970s and 1980s, but automobile ownership increased slightly. Food consumption probably increased slightly, but the evidence for this is very weak. .