In debates over Indian poverty trends in the 1990s, questions have been raised about the comparability of the quinquennial 50th and 55th rounds of the consumer expenditure survey (CES) carried out by the National Sample Survey Organisation in 1993-94 and 1999-2000. These focus on possible interference between alternative reference periods used to elicit expenditure data from sampled households. This paper resolves these concerns, using comparable consumer expenditure data from the employment-unemployment survey (EUS) of the 55th round, as well as data from four experimental rounds of the CES, conducted between the quinquennial larger scale 50th and 55th rounds. It shows that the size distributions of consumer expenditure from the 55th round CES are comparable to ones from the 50th round - subject to appropriate recalculation - and that there is accordingly unambiguous evidence that poverty in India declined in the 1990s, in all dimensions. Indeed, the average annual rate of reduction in the last six years of the 1990s is shown to have been higher than that between 1982 and 1993.