Generalised reverse projections of the population of Madras, 1866-1951: an evaluation of consistency

Ryuichi Kaneko, National Institute of Population and Social Security Research, Japan
Osamu Saito
Mihoko Takahama, Hitotsubashi University

This paper is an attempt to apply an inverse projection technique known as Generalized Inverse Projection (GIP) to data for a pre-independence Indian province. We have already tried to estimate annual series of TFR and e(0) for two pre-independence provinces, Madras and Punjab, by adopting the Brass relational fertility model, logit life-table system and the growth balance method for the period 1891-1951. In this paper, we evaluate these data and estimates in terms of the demographic consistency aiming at obtaining a further reliable estimate of the province’s population history. We applied GIP to population in Madras. In application of GIP, the major demographic variables are simultaneously estimated in a consistent manner with input data as a whole. We use the results to evaluate the data quality and the previous estimates. As a result, some inconsistencies in data and the previous estimates mainly due to incompleteness of the registration are detected.