1. Ian Savage and August Schupp (1997). Evaluating transit subsidies in Chicago. Journal of Public Transportation 1(2):93-117.
[Published Version]  [Manuscript Version]

This paper presents a model that calculates the social welfare benefits of using additional subsidy to reduce fare levels or improve service levels of public transit in Chicago. The model differentiates between the effects in peak and off-peak periods for both bus and rapid rail service.

The potential welfare benefits of transit improvements accrue to several different groups of people: (1) existing transit riders; (2) automobile users who decide to switch to public transit; (3) road users who benefit from reduced road congestion; and (4) people who undertake new trips. These benefits are compared with the welfare cost of the additional sales taxes that are used to fund transit subsidies in Chicago. Results of the analysis are that:

• Bus fares should be reduced, especially during off-peak and weekend periods;
• Rail fares are "acceptable" in that the marginal benefit of using subsidy to reduce fares is close to the excess burden of raising the subsidy;
• Bus service levels are broadly acceptable, except for the peak period where they are too high; and
• Rail service levels are too high at all times of the week, but especially in the peaks and on Sundays.

In general, it is more advantageous to use subsidy monies to reduce fares than improve service levels. Even if overall subsidy levels were not increased, society would be better off if service levels were reduced, and the money saved channeled into reductions in fares.

2. Ian Savage and Kenneth A. Small (2010). A comment on “Subsidization of Urban Public Transport and the Mohring Effect.” Journal of Transport Economics and Policy 44(3):373-380.
[Journal Website]  [Manuscript Version]

In an 2008 paper in the Journal of Transport Economics and Policy, Peran van Reeven argues that the "Mohring effect" - whereby subsidies increase ridership, the ridership increase engenders higher service frequencies, and the higher frequencies reduce transit riders cost by reducing the average waiting times at stops - is not relevant to the determination of transit subsidies because a profit-maximizing monopolist would supply frequencies that are the same as, or greater than, those that are socially optimal. Our analysis finds that his results depend on the reduction or elimination of the effect of fares on demand, causing optimal prices to be indeterminate within broad ranges. Consequently, van Reeven's model is an unsatisfactory tool for discussing subsidies in general, and the optimal combination of fare and frequency in particular.

3. Ian Savage (2010). The dynamics of fare and frequency choice in urban transit. Transportation Research A 44(10):815-829.
[Journal Website]  [Manuscript Version]

This paper investigates the choice of fare and service frequency by urban mass transit agencies. A more frequent service is costly to provide but is valued by riders due to shorter waiting times at stops, and faster operating speeds on less crowding vehicles. Empirical analyses in the 1980s found that service frequencies were too high in most of the cities studied. For a given budget constraint, social welfare could be improved by reducing service frequencies and using the money saved to lower fares. The cross-sectional nature of these analyses meant that researchers were unable to address the question of when the oversupply occurred. This paper seeks to answer that question by conducting a time series analysis of the bus operations of the Chicago Transit Authority from 1953 to 2005. The paper finds that it has always been the case that too much service frequency was provided at too high a fare. The imbalance between fares and service frequency became larger in the 1970s when the introduction of operating subsidies coincided with an increase in the unit cost of service provision.