Estimating Price Elasticities of Ferry Demand

Abstract

Washington State Ferries operates what is by far the largest ferry system in the United States. The ferry routes serve as the primary transportation link to several islands in Puget Sound. Customers include a mixture of commuters, nonwork travelers, and tourists who use the ferries as walk-on or drive-on passengers. The system experiences seasonal and time-of-day peaks that result in some routes operating at capacity for drive-on customers. As part of a recently completed long-range planning effort, Washington State Ferries in association with the state's transportation commission wanted to evaluate the impact of changes in fare policies—in particular, to analyze the effects of charging different fares at different times of day. Fare levels have changed several times over the past several years. The resulting elasticities corresponding to these fare increases were calculated and provide a reasonable foundation for estimating the effects of future fare changes. However, the elasticities do not reflect the effects of charging different fares by time of day. To estimate those elasticities, a stated preference survey was designed and administered to current drive-on customers. Data from the survey were used to estimate both segment-level discrete choice models for choice of alternative drive-on sailings, walk-on, or shifting to an alternative route or mode. Hierarchical Bayes estimation was used to develop models that reflect random heterogeneity in preferences and those models were used in a simulation model to estimate time-of-day fare elasticities. This paper describes the resulting fare elasticities and their implications for fare policy.