Econometric modeling of barge-freight rates on the Mississippi River

Brian M Wetzstein, Purdue University


The Mississippi River system is important for both domestic and international U.S. trade, and handles over 300 billion ton-miles of shipments annually. It plays a vital role in the global competitiveness of U.S. agricultural commodities, aids in the sustainability of alternative energy production, and provides competition to both rail and truck leading to efficient U.S. freight rates. Currently, over 60% of agricultural exports move on this inland waterway system from the U.S. Midwest to downstream export ports near New Orleans. In the face of increased international demand for United States exports, as well as the expansion of the Panama Canal, it is imperative that the efficiency of barge transportation on the Mississippi River is improved in the most cost effective manner possible. Thus the efficiency, profitability, and magnitude of the U.S. export market may be improved. This can directly spur economic growth in the form of increased GDP and job creation, as many non-farm sector jobs are related to agricultural exports. In order to investigate this, two different vector autoregressive models and two different spatial vector autoregressive models are developed. The dependent variable in all four models is the price of barge transportation. This study breaks the Mississippi into five distinct segments and explores the interaction among them. To populate these models, a variety of exogenous variables that affect barge rates are identified. For each variable, weekly data are collected from January 2003 to June 2014, resulting in 594 observations across five river segments. One way these models can improve the efficiency of barge transportation is through the prediction of barge rates. For each model, out-of-sample forecasts are constructed for the barge rate in each segment. Overall, the spatial vector autoregressive models produce the most reliable forecasts given a single out-of-sample forecast period and are statistically superior to naïve forecasts on the basis of a variety of forecast comparison tests. With a five-week forecast horizon, over a million dollars can be saved in transportation costs over a two-year out-of-sample period. In addition, it is possible to improve the efficiency of barge transportation with a close examination of the controlled depth of the river. Currently, the Army Corps maintains a nine foot deep and 300 foot wide navigable channel. This analysis finds that draft depths significantly affect barge rates and that the magnitude of the elasticity values is larger than that for lock delays. Further, only a small portion of the Army Corps’ budget is spend on dredging, which suggests that policy and budget changes can improve efficiency. This analysis produces a dollar value for the cost to increase the channel depth of the segments under study by one foot and examines the savings in transportation costs this will produce given the associated reduction in barge rates. Environmental factors, climate change, and points of future research are also identified.




Foster, Purdue University.

Subject Area

Agricultural economics|Transportation planning

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server