Forecasting U.S. Automobile Sales

The purpose of this project is to research on the issue of auto sale in United State, which is critical to the automobile industry as well as overall economy. The American automobile market is affected by various factors, from both internal and external environment. In order to figure out the correlation among factors, we research on its historical performance in relation to many factors that may affect the automobile industry. Data about automobile sales and many economic and demographic variables are collected from a variety of sources. Automobile sales within U.S. are the dependent variable and the rest are independent variables. By using stepwise method in SPSS, the significant variables are selected and the first regression model is developed based on these variables. Next, we add a lag to all and some independent variables separately and come up with the second and third regression models. By comparing the adjusted R-square, we finally choose the optimized regression model to forecast U.S. automobile sales. The results indicate that some of the factors which significantly reflect the internal and external impact on the automobile market. By implementing historical data of each factor to the model, it can forecast the auto sale in U.S. domestic market. The results in this study could help the automakers better understand the domestic auto market.