Many simplistic comparisons have recently been made between the NSS estimates of household expenditure and the independent data on private consumption from the national accounting system (NAS). This paper attempts to (a) investigate the scientific rationale behind these comparisons, (b) indicate the weaknesses and strengths of the external validator data set as a touchstone to assess the reliability of the NSS estimates, (c) assess the degree of cross validation, or the lack of it, between the two data sets relating to two years in the 1970s, and (d) draw some lessons for improving the two data sets as well as caution against their incorrect uses. Two supplementary notes, one dealing with standard errors of the NSS estimates of consumer expenditure and the other with the possibilities of non-sampling biases in the NSS budget estimates of cereal consumption, are provided in the appendices.