UCLA Luskin School of Public Affairs

The federal government, along with many states, has adopted policies favoring the provision of public transit by the private sector. During the 1980's, this turn to contracting to halt rising operating deficits prompted several studies into the impact of contracting on operating efficiencies.

Most research found that service contracting saves 10 to 60 percent over publicly operated services. However, no research has yet examined the long-term cost trends of private contracting vis-a-vis public operations. The evaluations done to date often make inappropriate comparisons between small single mode private carriers and large multi-service transit authorities with greater political and social obligations. As a result the findings from these studies are certain to show dramatic savings, yet do not address the underlying dynamics driving transit costs such as political pressures to provide service.

This study examined cost efficiency trends for 142 transit operators providing fixed-route bus transit between 1989 and 1993. This analysis produced no evidence that fully contracted operations cost less per revenue hour than publicly operated services doing no contracting. Vehicle and driver scheduling inefficiencies were found to contribute the most unit costs. Estimated elasticities indicate that a 10 percent reduction in vehicle scheduling inefficiency may produce a 19 percent improvement in cost efficiency. A 10 percent improvement in operator scheduling efficiency shows a 6 percent reduction in operating costs per revenue hour. These findings indicate that transit service contracting may not produce cost savings over the long-term and that strategies of decentralization and changes in the craft structure for labor may be more appropriate ways for relieving the fiscal crisis of public transit.

Over the course ofthis century, public transit systems in the U.S. have lost most ofthe market share ofmetropolitan travel to private vehicles. The two principal markets that remain for public transit systems are downtown commuters and transit dependents- people who are too young, too old. too poor, or physically unable to drive. Despite thefact that transit dependents are the steadiest customers for most public transit systems. transit policy has tended to focus on recapturing lost markets through expanded suburban bus, express bus, andfued rail systems. Such efforts have collectively proven expensive and only marginally effective. At the same time, comparatively less attention and fewer resources tend to be devoted to improving well-patronized transit service in low­ income, central-city areas serving a high proportion of transit dependents. This paper explores this issue through an examination of both the evolving demographics ofpublic transit ridership, and the reasons for shifts in transit policies toward attracting automobile users onto buses and trains. We conclude that the growing dissonance between the quality of service provided to

inner-city residents who depend on local buses and the level of public resources being spent to attract new transit riders is both economically inefficient and socially inequitable. In light of this. we propose that transportation planners concerned with social justice (and economic efficiency) should re-examine current public transitpolicies andplans.

The difficulty of collecting social conflict data has caused its study to focus on recent events predominantly in the West. This paper shows how to use artificial intelligence to generate social conflict data regardless of language, location, or date. Digitizing tertiary books, lightly cleaning their text representation, and submitting that text to a large language model (LLM) produces accurate date, location, and event descriptions. These capabilities and results are demonstrated with books on Latin America after 1492, Imperial Russia, and Tokugawa Japan. Using LLMs requires a larger fixed cost than working a team of research assistants, but it produces results more quickly for most sources. It is less accurate for the Tokugawa Japan source; whether it or human translation is cheaper depends on the cost of correcting the LLM’s work. These results represent the floor of accuracy for artificial intelligence, suggesting researchers should soon use it for creating social conflict data from tertiary sources.