Quantifying the Technical Efficiency of Canadian Paratransit Systems Using Data Envelopment Analysis Method
Paratransit service operators in Canada are under increasing pressure to improve the operational productivity of their services due to increased demand and tightening financial constraints. To achieve this, Paratransit operators need to know their performance as compared to peer systems and the best practices within the industry. This will enable each operator to identify where and how much improvement should be made in order to be on a par with the industry?s best practices. Little research effort, however, has been devoted to the issue of how to measure and compare paratransit efficiency in a consistent and systematic manner. <br /><br /> This research focuses on evaluating the level of efficiency of individual paratransit systems in Canada with the specific objective of identifying the most efficient service agencies and the sources of their efficiency. By identifying the most efficient systems along with the influencing factors, it is possible that new service policies and management and operational strategies could be developed for improved resource utilization and quality of services. To achieve this objective, this research applies the analysis methodology called Data Envelopment Analysis (DEA) approach which is a mathematical programming based technique for determining the efficiency of individual systems as compared their peers involving multiple performance measures. Annual operating data from Canadian Urban Transit Association for Canadian paratransit systems of year 2001, 2002 and 2003 are used in this analysis. Regression analysis is performed to identify the possible relationship between the efficiency of a paratransit system and some measurable operating, managerial and other factors which could have an impact on the performance of paratransit systems. The regression analysis also allows for the calculation of confidence intervals and bias for the efficiency scores in order to assess their precision.