The U.S. economy depends heavily on the trucking industry as it moves 70% of the entire nation's freight. With the inclusion of $295 billion in truck trade with Canada and $195.6 billion in truck trade with Mexico in 2007, it is apparent that any disruption in truck traffic will lead to rapid economic instability (ATA Releases: American Trucking Trends 2008 - 2009, 2008). Yet, the critical nature of the trucking industry comes at a societal price. Indeed, undesirable outputs, e.g., truck crashes and associated injuries and fatalities, have very significant economic and human consequences.

This dissertation uses Data Envelopment Analysis (DEA) to investigate the impact of undesirable outputs on the productivity of the motor carrier industry during the years 1999-2003. Previous DEA studies at the firm level have focused on the relationship between inputs and desirable outputs. The proposed approach in this dissertation simultaneously considers both the positive and negative outputs.

This dissertation addresses two key problems with the DEA analysis technique previously identified by Yang and Pollit (2009): i.e., failure to take into consideration undesirable outputs and the failure to assess the impact of exogenous variables on the DEA scores of individual firms. As a result, this study will provide a new perspective into the productivity of U.S. motor carriers by incorporating both of these considerations into a more comprehensive DEA analysis. It will also provide opportunities to evaluate how individual firms might change their mix of inputs in order to simultaneously maximize desirable outputs and minimize undesirable ones.