Forecasting Australian marriage rates

Area of Science:

• Demography
• Econometrics
• Time Series Analysis

Background:

• Accurate demographic forecasting is crucial for policy and planning.
• Modelling demographic variables requires careful selection of forecasting equations.
• Previous studies have utilized regression methods for demographic forecasting.

Purpose of the Study:

• To evaluate empirical model selection procedures for demographic forecasting.
• To compare the forecasting performance of final equation models with ARMA models for Australian marriage rates.
• To examine the properties of ARMA model forecasts in a demographic context.

Main Methods:

• Application of two empirical model selection procedures.
• Development of final form forecasting equations for Australian marriage rates.
• Comparison of post-sample forecast performance against univariate ARMA-type models.
• Examination of ARMA model forecast properties.

Main Results:

• The study identified final form forecasting equations for Australian marriage rates.
• Autoregressive Moving Average (ARMA) models demonstrated superior post-sample forecast performance.
• ARMA models were preferred over regression-based approaches for forecasting marriage rates in this instance.

Conclusions:

• ARMA models are recommended for forecasting Australian marriage rates.
• The findings suggest a preference for time series models over traditional regression methods for this demographic variable.
• Further examination of ARMA model properties provides insights into demographic forecasting strategies.