Growth rate of a population following logistic curve has a definite pattern: rising from an extremely low level to a maximum level sometime and declining after that, gradually reaching zero. Logistic model has also been found to be of immense use in predicting subpopulations (Leach, 1981) because it provides a working model of the “mechanism of self correction”. UN Manual XIII (UN, 1974) showed that a constant urban-rural growth difference (URGD) leads to logistic growth of the degree of urbanization (i.e., percent urban).

Subsequently, URGD method was used for projections of ratio of urban population to total population, and even other types of ratios such as ratios of populations of cities to total urban population, ratios of labour force to total population and school enrolment rates. Thus urban population can be predicted by multiplying projections of total population by projected ratios of urban population to total population. Projections of total population are obtained using component method (Smith, 1992).

ESTIMATION OF MORTALITY AND FERTILITY

In 1950's when the studies of population started in both developed and developing countries, lack of data on the vital rates was a big problem for most countries. This was the best period for development of mathematical models in population studies. Mathematicians and statisticians developed models to estimate birth rate, death rate, total fertility rate and life expectancy using incomplete and unreliable data (Shryock et al., 1971; Roy, 1987). The main issue was: can we get working estimates of fertility and mortality using census growth rate, age distribution of population, and open and closed birth intervals found from survey data? Mathematical models helped in this greatly. For example, when reliable data on total fertility rate (TFR) did not exist but data on parity by age was available from surveys, using an empirical relationship, Coale and Demeny suggested the following approximate formula: