Estimating Comparable Poverty Counts from Incomparable Surveys: Measuring Poverty in India

摘要: We develop a procedure to estimate poverty counts in India from the 55th Round of Na- tional Sample Survey (NSS), large household survey run 1999-2000. The evidence suggests that change design caused reports on expenditure an extent it is impossible, without adjustments, compare estimates this with those obtained previous NSS Rounds. More generally, paper addresses problem comparing distribution variable across differently designed surveys, when different causes respondents' about be incomparable surveys. proposed requires only existence set auxiliary variables whose are not affected by design, and relation main interest stable estimator, instead, does require specific functional form assumptions between variable. In context data, we identify systematically changes implemented provide stability over time per capita total conditional these variables. describe experiment evaluate performance showing provides satisfactory results, both estimation density expenditure. Finally, use our estimator calculate adjusted for using data 1999-2000 Survey. results show sharp reduction nineties, even if rural areas as implied unadjusted figures.