Evaluation of Poverty Indicators
T-PEMD-88-1, Mar 22, 1988
GAO discussed the Census Bureau's experimental methods for valuing noncash benefits as part of measuring poverty in the United States, specifically whether it assigned a cash value to a benefit based on: (1) what it would cost to buy the benefit (the market value method); (2) what it was worth to the recipient (the recipient value method); or (3) what percentage of its budget a poor family would spend on the benefit (the poverty budget share method). GAO found that: (1) both the market value and recipient value methods revealed substantial conceptual and methodological problems, especially for valuing medical care; (2) the market value method underestimated the extent of poverty, affected the poverty rates of subgroups in varying ways, disrupted the poverty gap distribution, and assigned high benefit levels compared to its composite analysis; and (3) the recipient value method was likely to reclassify blacks, the elderly, persons in families headed by women, unrelated individuals, single women, and Hispanics as poor when medical benefits were decreased beyond 10 percent and as nonpoor when they increased beyond 10 percent. GAO believes that further empirical analysis and more data are needed to confirm the extent of the problems and to identify problems not critiqued.