Stochastic forecast of fertility

Future population trends are of interest to a wide range of analysts, including policymakers, scientists, and planners in industry and government. Global and national trends in population size are used to estimate the future demand for food, water, and energy, as well as the environmental impact of rising consumption of natural resources. A new method is proposed for forecasting age-specific fertility rates observed over time. This approach can be summarized as follows. (1) Fit the suitable family of distributions for modelling the fertility rates. The age-specific fertility pattern has a typical shape through years. In order to describe this shape a number of distributions have been proposed. Most commonly used distributions are beta distribution, gamma distribution, Hadwiger distribution, mixture of beta distributions and mixture of Hadwiger distributions. (2) Estimate the parameters of distribution. (3) Forecast the parameters using the time series model. (4) Use the forecast parameters to forecast the model for age-specific fertility rates by one-year age groups.The analysis is based on the data of the Estonian Statistics Office. As the result of historical circumstances we cannot use a very long period of time. Therefore we are going to observe only period post re-independence which is 1991-2009.