Forecasting automobile sales: the Peña-Box approach

Abstract

As a response to growing concerns regarding the call for clean energy and its impact on future automobile sales, this study uses a classical factor model and the Peña-Box model to examine the contemporary and time-varying relationships of different brands/ models of cars in Taiwan between 2003 and 2007. In this paper, we demonstrate the complementary characteristics of these two analytical and forecasting methods. The results confirm that these two models can derive equally important but different information from the same time series data. Furthermore, the models are a useful marketing tool for discovering the current preferences of car purchasers, as well as their preference changes over time.