EVALUATION OF A NEW SIMPLIFIED POPULATION PROJECTION MODEL: A CASE STUDY OF LOCAL GOVERNMENT AREA PROJECTIONS IN TASMANIA

Preparing local area population projections with state-of-the-art demographic models can be a challenging, time-consuming and costly task. Alternative simpler models can produce projections quickly and easily, but at the cost of less output detail, less flexibility in creating scenarios, and sometimes lower accuracy. This paper presents an evaluation of a new modelling approach which blends the conceptual sophistication of state-of-the-art cohort-component models with the low data requirements of simple models. A key feature is that no locally-specific fertility, mortality, or migration input data is necessary. The new model is tested by producing ‘projections’ of local government area populations by age and sex in Tasmania over recent periods, with the results then compared to actual populations. The model is shown to produce reasonably accurate projections, and out-perform a simple benchmark model. The strengths and weaknesses of the new approach are discussed.