Abstract

The downward spiral frequently experienced by transit companies when they increase fares to make up for declining demand is well known. It is characteristic of an increasing-returns-to-scale industry where consumer inputs are essential to the production process. This paper formally incorporates both agency and consumer inputs in a cost framework and explores the short- and long-run costs of bus services. In so doing it is possible to trade off both consumer and agency costs, to arrive at optimal service levels. It is these service levels, with the mode "performing at its best," that form the basis of cost comparisons between transportation alternatives. We find that off-peak frequencies should be considerably shorter than those typically prevailing on bus transit systems, and that marginal cost pricing would require a zero fare in the off-peak and a fare of up to 20 cents in the peak. This suggests that current rates might be continued in the peak and abolished in the off-peak. A subsidy to maintain this scheme would be about 10% of the average cost of service when operation is in long-run equilibrium (optimized). In the realm of cost comparisons between alternative modes, options based upon the car are found to be among the cheapest for trips to the CBD except in the peak. In terms of public modes, dial-a-ride systems seem to have some economic justification for low density neighborhoods as a feeder to linehaul buses operating on freeways, but at high densities integrated bus service appears to be far more viable. The latter is markedly cheaper than a feeder bus + rapid rail alternative. Finally, there seems to be little economic justification for the current interest in non-standard buses for transit service.