Methods for projecting fertility

Fertility forecasting in the case of Brazil constitutes a great challenge due to the regional heterogeneity of its fertility transition, making the application of the known projection techniques difficult. Thus, the paper presents a methodology that allows to forecast the pattern and level of fertility by defining scenarios for a small geographical unit based on the fertility behavior of the total population. In this case, Brazilian Federate States (BFS) and the whole country as a unit. We use National Household Surveys, Demographic Censuses (2000 and 2010) and vital statistics where data are reliable. We assume that the trend of reproductive behavior outlined for the country as a whole is a transition process to be experienced by all BFS, differing only by the timing it occurs. From this assumption it is possible to identify –using the more recent BFS’ TFR as a first parameter– the timing of its corresponding fertility transition. Then, using interpolation procedures we replicate the national transition experienced by each ASFR. The robustness of this technique is given by the coincidence of the sum of births generated by the FBS’s ASFR and the total births generated by the ASFR defined for the whole country.

Given the fact that a satisfactory estimate of cohort fertility depends crucially on an accurate prediction of the future trend of period quantum, this paper shows that one can utilize available fertility data
to disclose some useful information about that trend so as to effectively correct the prediction bias occuring under the no-quantum-change anticipation. Specifically, we extract clues about both the slope and the change of slope in current quantum movements, and then exhibit a very high correlation between the slope of period quantum and the prediction bias which comes from a large number of experiments
by fully utilizing the existing data from Canada, the U.S., and 23 European countries, As a result, the prediction bias can be significantly corrected based on this relationship so that a satisfactory estimate of cohort fertility is thus obtained.

The Stable Bounded Model of Fertility and Time, is a mathematical model that provides theoretical elements and specific applications in order to minimize the possible errors that can occur in the process of projecting total fertility rate through time. The model generates scientific demographic knowledge to estimate the value of the stability of the total fertility rate (TFR), by the method of Quantities Change and Converging Autoregressive Processes. It also offers elements to find the function that best fits the observed data of TFR and stabilizing in the future in the estimated value by the method of Separable Differential. The model is a new methodology for projecting fertility, which provides reliable estimates in the short, medium and long term and different solutions to the problems of the current methodology used to project the level of fertility. Model is applied in countries where TFR stability has already happened, in order to compare the prediction of stability of model with the real value observed in these countries. Also, is applied in countries where has not happened, in order to project the TFR until 2050.

For the new round of IIASA/Oxford education projections, we present a new fertility scenario making process that attempts to overcome limitations of regular assumption making. Firstly, we include a large number of international experts to feed into the scenario making process by answering a questionnaire to identify main drivers of fertility, and secondly, a group of meta-expert evaluates results from the questionnaire. The construction of the scenarios consists of a three staged modeling approach. First, we estimate a model, using a country’s level and decrease of fertility during the past five-year period and compare it to countries with similar characteristics since 1970. Second, we estimate expected decrease of fertility by employing information, gathered from the fertility questionnaire. And third, numerical point estimates, supplied by the meta-experts, are utilized to estimate future fertility decline. Combining the information from three qualitatively very different sources, we are able to provide a new set of fertility assumptions to feed into the IIASA/Oxford education projections. This paper discusses a new assumption making approach for countries in today’s high fertility world and compares the differences in methods and results to the Bayesian projection methodology introduced by the United Nations.

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.

In this paper we aim to study the fertility prospects of Brazil major regions based on a cohort analysis, starting from the mid-1960s. Using micro-census data from 1980 to 2010 it is reconstructed the fertility history of women in five macro regions of the country, namely: North, Northeast, South, Southeast and Midwest. Based on the complete birth history, we reconstruct cohort fertility and afterwards, we apply two methods to analyze the past, present and future trends in Brazilian fertility. First, it is applied a basic benchmark cohort model in order to understand the past and present progress of fertility. For the future prospects, we apply a New Cohort Fertility Forecasts, developed by Myrskylä et al. (2012). As results, we see that there are clear regional differences in cohort fertility. Even in cohort perspective, the levels of Brazilian cohort fertility are still below replacement levels and very concentrated at the young ages of reproductive span. However, in the more developed South and Southeast parts of the country there are signs of fertility postponement. This will result in further decline of fertility in the near future.

We propose a new Bayesian approach to forecasting cohort childlessness. Combining historical and contemporary data from the Human Fertility Database (HFD), we estimate a posterior distribution for the Lexis surface of age-specific first-birth rates for US cohorts born 1950-1992. Past rates on this surface are known with high precision from the HFD, while future rates must be forecast. Our approach combines estimation of past and future rates in a single model, using historical HFD data to build priors and thus identify likely (and unlikely) age and time patterns across Lexis surfaces. The resulting forecast of first-birth rates and cohort childlessness automatically includes uncertainty estimates. Among many other results, our forecast indicates that US childlessness, which is currently falling slightly, will reach a minimum for women born in the 1970s, and will almost certainly be higher for those born in the 1980s.

This paper estimates how far current fertility levels and rates of population growth in Kenya can be reduced purely by relying on family planning or – more precisely – the elimination of unwanted fertility. Using a similar approach to Bongaarts decomposition model, the paper projects the population, starting in 2010, using the Age Specific Fertility rates of each respective 5-year period, but adjusted by a factor. By 2050, the population implied by this scenario would be 104.1 million, compared to 96.9 million under the UNPD Medium projection and 94.6 million under the previous scenario with uniform reduction of the ASFRs. By 2070, the Perfect Reproductive Health scenario would imply a population of 168.6 million, compared to the 127.3 million projected by the UN Population Division. If maternal mortality is completely eliminated and a further reduction of 50% in child mortality is assumed, the former number rises to 176.1 million. Although the immediate attainment of perfect reproductive health would lower population growth rates in the short run, such improvements – in the absence of changes in fertility preferences - would soon exhaust their potential, resulting in long-term population growth rates in the order of 2-2.5%.

Total fertility rate as a social parameter affects by different economic, social, cultural and institutional factors. Inglehart’s theory of cultural change explains and theorizes social changes sourced economically within a time process.
The main goal of this study is to examine Inglehart’s theory of cultural change in terms of rising trance material values, adopted by effect of economic indices on total fertility rate, by modeling and expecting total fertility rate up to 2025, based on four economic variables: GDP growth, gross saving, consumption price index and GDP per capita growth, and four social indices including: life expectancy at birth, total fertility rate, literacy level of total population and +25. Within three economic scenarios: continuing current trend, subsidize law and achieving 2025 vision goals.
The used data of socio-economic indices were worked out from Statistics of World Bank and Statistical Center of Iran.
Methodology of this study was based on Neurotic Net Work ( that named GMDH algorithm).In this study, we put selected variables affected on fertility change in Iran in the model and then predicted total fertility rate for 5 years periods up to 2025.
Based on three scenarios total fertility rate will reach to 1.34, 1.33 and 1.29 in 2025.

Using empirical fertility rates and population distributions, we study comparative contributions to births’ prediction errors of choices for the fertility model and of the approximation errors of three main fertility indicators (the total fertility, the mean and the standard deviation of age at birth, respectively: TFR, MAB, SDAB). Agreeing with theories of dynamic populations, we find high importance of accuracy of TFR and MAB. Yet, the role is limited in population projections of the estimates of SDAB and of the choice of the fertility model form. More attention may be paid in population projections to working out (interdependent) scenarios for TFR and MAB, while relaxing complexity of other aspects of fertility projection models. Our results suggest widening the uncertainty range for TFR in cases when the MAB projections are based on regressions on TFR or other simplified assumptions.