Thursday, September 17, 2015

Forecasting Vehicle Kilometers Traveled: Estimating an Autoregressive Integrated Moving Average Model (ARIMA) with Exogenous Variables

Recent working paper:

Nazemi, M., Shafiei, M.S., Saberi, M., Sarvi, M. (2015) Forecasting Vehicle Kilometers Traveled: Estimating an Autoregressive Integrated Moving Average Model (ARIMA) with Exogenous Variables. [Working Paper]
Vehicle kilometers traveled (VKT or VMT) is a key variable in long-term transportation planning and policy making, especially for road infrastructure investment. In this paper, we estimate an Autoregressive Integrated Moving Average (ARIMA) model with exogenous variables to forecast VKT in Australia. The estimated model produces forecasts based on lagged values in the time series and the errors made by previous predictions, which typically allows the model to rapidly adjust for sudden changes in trend, and at the same time includes exogenous variables resulting in more accurate forecasts. We found that the effects of unemployment rate, fuel price over Gross Domestic Product (GDP) per capita, and Consumer Price Index (CPI) on VKT are all statistically significant and negative. More importantly, number of vehicles per capita appears to be highly statistically significant with positive impact on VKT. We postulate that policies or behavioral changes towards less car ownership are more likely to have greater impact on VKT rather than fluctuations in fuel price, unemployment rate, or individual’s purchasing power.

No comments:

Post a Comment