This paper details the results of a study of a 15 cent across the board increase in fares carried out in July of 1981, by the southeastern Michigan transportation authority (SEMTA). Based on behavioural hypotheses, data from the same riders were gathered both before and after the fare increase (i.e. a panel); econometric methods tailored to these data were used. A limited dependent variable model ("TOBIT") was developed to account statistically for the fact that some people stopped using transit entirely (i.e. had zero trips after the fare increase). The results indicate that sensitivities to the fare increase among riders in this sample vary noticeably across types of service. In particular, transit trip-makers in the off-peak appear to be nearly twice as sensitive as those in the peak periods. It is clear that a wide variety of factors influence the use of bus transit. Even if transit service variables (e.g. headways) are virtually constant during the period of study (as was true in the SEMTA system), many non-transit factors (e.g. whether a person works, location of job, location of home) change continually for enough riders to make a difference. Conclusions drawn about riders' sensitivities to fare increases which do not account for changes in non-transit factors are likely to be erroneous. Because of the special nature of the data (from on-board surveys in the before case) and the several layers of statistical and theoretical concerns which arose, a number of issues were identified for further research.(a) for the covering abstract of the conference see IRRD 276520.
Abstract