Summary: Pursuant to a congressional request, GAO discussed on the development of a methodology for evaluating proposed changes to poverty indicators and thresholds. GAO found that: (1) the valuation method for medical benefits assigned an average benefit level to all program participants, and the extensive and expensive services provided to those who were in the terminal period of their lives were credited to the well-being of all participants; (2) the market-value technique may overvalue the worth of a benefit; (3) the cell-matching procedure, used to tabulate the expenditures of selected groups, may risk selectivity bias by exaggerating larger or smaller expenditures; and (4) underreporting income may lead to considerable error in the poverty rates. GAO also found that: (1) there has been little direct, in-depth technical analysis of the three methods for valuation of noncash benefits, even though a wide array of technical concerns have been raised about them; (2) the Bureau of the Census has made little effort to test its methods; (3) the differences in valuation rates depend on what factors are included in the estimation procedures; (4) most of the larger differences between poverty reduction estimates were accounted for by the inclusion of medical transfers; (5) the Bureau's published work on valuing noncash benefits did not directly address the concerns it had had since 1982, or earlier; (6) the work done on the Survey of Income and Program Participation has been devoted to achieving comprehensiveness and quality with respect to data on cash income and program participation; and (7) further empirical analysis and greater information distribution are needed to confirm the extent to which problems have not been critiqued in detail.
