Growth Forecasting of Vehicle Miles of Travel at County and Statewide Levels

Abstract

Growth forecasting of vehicle miles of travel (VMT) plays an important role in achieving a variety of goals and objectives in transportation planning, air quality planning and conformity analyses, and funding allocation for highway improvement programs. States have used a variety of methods to forecast statewide VMT growth, including historical growth factors, travel survey–based methods, and fuel consumption–based methods. The growth factor method appears to be the most popular technique. However, this method has a major limitation in that growth factors do not respond to changing socioeconomic conditions over time. In this study, statistical models were developed to forecast VMT growth at county, county aggregation, and statewide levels; these incorporated changing socioeconomic trends and forecasts for the future, as well as transportation system supply variables. During the modeling process, a variety of variables, different forms of these variables, and indices, including socioeconomics (population, households, employment, income), density and mix variables, lagged variables, difference form, and county dummy variables, were tested. Different types of regressions and model specifications were tested and evaluated for statistical validity and forecasting validity at county, county aggregation, and statewide levels. A selected set of models was used to forecast VMT for planning horizon years 2010, 2020, and 2030. Forecasting results were evaluated on the basis of forecasting trend versus historical trend, growth magnitude at state and county levels, and spatial patterns of growth. Finally, a preferred forecasting system was recommended for implementation. This study, a Pennsylvania case study, offers a VMT forecasting methodology that is low in cost and transferable to other states.