Semester

Spring

Date of Graduation

2013

Document Type

Dissertation

Degree Type

PhD

College

Statler College of Engineering and Mineral Resources

Department

Civil and Environmental Engineering

Committee Chair

Avinash Unnikrishnan

Committee Co-Chair

Udaya B Halabe

Committee Member

Donald Lacombe

Committee Member

David R. Martinelli

Committee Member

Santiago Pinto.

Abstract

Truckload (TL) pricing is a major factor that influences the manufacturing and retail costs of products. In the U.S., trucks accounts for more than 90% of freight shipped based on value, and it is expected to grow in the following years. TL price setting is a very complex task for logistic companies as it depends on a number of factors including the logistics carriers' business strategies and other social and economic variables. Understanding TL patterns across the U.S. is important not only for logistic companies, but also for policy makers. TL prices are commonly provided on a dollar per mile rate. Thus the total transportation costs on a route will be the product of the truckload price rate and the distance. More accurate prediction of TL price will enable logistic companies to develop more optimal strategies to operate their transportation activity across destinations and effectively allocate resources on potential demand locations. Freight and economic policy makers will also be able to use this information to explore different potential economic scenarios.;This research analyses private data sets (TL rates), and publicly available data such as diesel cost, unemployment, wages, population, and gross state product to understand trends in TL prices. TL rates are evaluated through exploratory and visualization techniques to obtain useful insights. Time series analysis (TSA) and spatial econometric analysis (SEA) are conducted for forecasting TL prices. TSA provides with a general model based on time and delivery distance between origin and destination. Spatial econometric panel models incorporate the spatial dependency, being used for drawing inferences across space, and also for forecasting TL prices. Results indicate that TL prices are closely associated with unemployment, which links the consumer spending with transportation cost. Diesel cost has not impacted TL prices significantly during the last years, as is evidenced in the TSA and SEA. Moreover, in low demand condition such as high unemployment, carriers are likely to serve larger delivery distance in order to reduce TL prices, which impact TL prices in neighboring locations. Increasing the delivery distance by 1.00% was found to reduce the price in dollar-per-mile by about -0.25%, and raise prices in neighboring locations by about +0.05%. Similarly, 1% increase in unemployment rate was found to reduce prices by about -0.30% and increase prices in neighboring locations by about +0.06%. Forecasting models indicate accurate TL price values, with MAPE values less than 10% for the TSA model for estimating an overall monthly price in the U.S.; and less than 20% for the SEA that consider spatial dependence for estimating a yearly price at each U.S. state. This research represents a benchmark in the analysis of freight prices, providing useful insights, identifying significant variables impacting TL prices, and potential methodologies for forecasting truckload prices.

Download

Share

COinS