Past studies of economies of scale in transit have tended to treat the transit industry as a set of similar agencies that have the same cost structure, and have conducted an analysis for the entire range of agency size using a single regression function. This may have caused problems in the interpretation of analysis results because agencies of different sizes that may have different cost structures are treated as a homogeneous data set. In this paper, I examine the issues of economies of scale in the provision of fixed-route bus transit service in the USA, using level of contracting as a variable to classify agencies into three different size groups. In this analysis, using a significantly larger pooled data set (compared to previous studies) constructed from the National Transit Database of the US Federal Transit Administration from 1992 to 2000, 1 found that agencies with different levels of contracting exhibit very different relationships between cost per vehicle hour and agency size. Applying the observed range of agency size, I also found diseconomies of scale for all agency Sizes with all levels of contracting, even when I utilized a quadratic function for the regression equation. While the analysis is limited since the model did not control for many other explanatory variables, the findings suggest that the level of contracting is possibly an important variable, among many attributes of transit systems, with which to classify agencies into different size groups, each of which has a different transit cost structure. (C) 2008 Elsevier Ltd. All rights reserved.
