Comparing the Efficiency of Public Transportation Subunits Using Data Envelopment Analysis
چکیده
This article discusses the need for a performance measure that compares the efficiencies of subunits within a transportation organization, reflects the diversity of inputs and outputs, and is objective and consistent. The study presents a method for developing such a performance indicator, and illustrates its use with an application to the park-and-ride lots of the Chicago Transit Authority. The proposed method applies Data Envelopment Analysis supplemented by Stochastic Frontier Analysis to estimate efficiency scores for each subunit. The research demonstrates how the scores can provide objective and valid indicators of each subunit’s efficiency, while accounting for key goals and values of internal and external stakeholders. The scores can be practically applied by a transit agency to identify subunit inefficiencies, and, as demonstrated by several brief case studies, this information can be used as the basis for changes that will improve both subunit and system performance. Introduction Due to shortages of public funds and expanding societal needs, maintaining and improving the performance of public transportation systems are critical for future operations (Kittelson et al. 2003; Sulek and Lind 2000). If public transportation Journal of Public Transportation, Vol. 10, No. 2, 2007 2 is not as efficient as it could be, it provides less service than desirable or requires taxpayers and riders to pay more than necessary. Improving the efficiency of a transit system’s subunits is one way to increase overall efficiency. Of course, maximizing subunit efficiency does not necessarily maximize system efficiency. However, overall system efficiency can be increased by correctly identifying subunit inefficiencies, and then improving subunit performance with changes that are consistent with system structures, goals, and constraints. For example, the efficiency of subcontracted service providers could be compared, as could the efficiency of individual bus routes, different rail lines, park-and-ride lots, rail stations, garages, and paratransit operations. If some of the subunits performing a given type of activity are identified as relatively inefficient compared with others performing the same activity, then management can take action to improve the least efficient ones, thereby improving overall system performance. The challenge lies in identifying and quantifying objective measures that reflect the multiple outputs and inputs common in public transportation. This article describes a procedure for comparing subunit efficiency, and demonstrates its application to the Chicago Transit Authority (CTA) park-and-ride lots. Park-and-ride facilities are a strategic component of urban mass transportation systems, effectively extending the service area and attracting riders who may not have otherwise used transit. However, with the high cost of construction, land acquisition, and maintenance of parking facilities, a misplaced or underutilitized lot can quickly consume significant resources with little promise for return on investment. The CTA has 7 park-and-ride lots, which not only provide parking for heavy rail passengers but also generate more than $ million annually in net operating income. Similarly, the Chicago commuter rail lines (operated by Metra) provide 68 lots, and Metra has been significantly increasing its lots’ capacities to attract more riders (Ferguson 2000). Indeed, park-and-ride facilities have become an integral part of almost all medium and large urban transportation systems in North America (Turnbull et al. 2004). This discussion addresses problems in developing valid efficiency measures that ( ) simultaneously incorporate all resource inputs and all desired outputs, (2) adjust for the influences of environmental variables, (3) are objective and can be rationally interpreted, and (4) provide consistent measurements with which to compare subunit performance. We describe an innovative solution to the problems—Data Envelopment Analysis (DEA). We identify the DEA model used, and the inputs and outputs that are most valued by the CTA. Using these inputs and Comparing the Efficiency of Public Transportation 3 outputs for the 6 lots for which we have data, we apply DEA. We use Stochastic Frontier Analysis to adjust the lots’ DEA efficiency scores for environmental factors. Finally, we present several brief case studies of lots that have been identified as inefficient by DEA. Problems with Measuring Efficiency of Subunits within a Transit Organization In economics, efficiency (or more specifically, technical efficiency) is measured by the ratio of output to input (Cooper, Seiford, and Zhu 2004; Färe, Grosskopf, and Lovell 994). In public transportation, multiple outputs are produced by multiple inputs, so different efficiency rankings may occur depending on the specific output/input ratios chosen for analysis. For example, although the primary goal of CTA management is to provide parking for heavy rail passengers, the profit derived from these lots also is an important output according to the agency. Thus, the key outputs are ( ) number of parked cars, as a proxy for number of passenger trips and (2) parking revenues. Key inputs are ( ) number of parking spaces and (2) operating expenses. Of course, other outputs and inputs could be added, but these are key to the CTA and are used to illustrate the process. Four different efficiency indicators can be constructed from these two outputs and two inputs: cars/expenses, revenue/expenses, revenue/capacity, and cars/ capacity. Table shows the actual performance of four CTA lots on these four ratios. Table 1. Comparative Efficiency of Four CTA Park-and-Ride Lots, First Quarter 2005 Lot Cars/ Revenue/ Revenue/ Cars/ Expenses Expenses Capacity Capacity A 3.05 4.25 . 8 0.85