This study addresses two significant limitations in the literature on cross-country expenditure comparisons: (a) treatment of all countries, large and small, as single entities with no spatial differences inside the countries, and (b) use of Divisia price indices, rather than Rank 3 preference-based exact price indices, in the expenditure comparisons. This paper compares alternative preference consistent methods for estimating spatial price differences in a large heterogeneous country, namely India, that are benchmarked against the spatial prices generated by the Laspyeres and Tornqvist price indices. Unlike the use of conventional price indices, the use of demand-systems-based methods allows the incorporation of price-induced substitution effects between items. The paper illustrates the usefulness of the methodology by using the exact spatial price indices, in conjunction with the inequality-sensitive welfare measure due to Sen, to rank the Indian states and examine changes in ranking during one of the most significant periods in independent India. The results have methodological and empirical implications that extend far beyond India.