In this presentation we describe an individual-level bicycle ownership model that was estimated as part of a new activity-based modeling system for Portland, Oregon. The bicycle ownership model is applied upstream of other activity and travel model components, conditioning mode choice sets to explicitly recognize that bicycling may not be available to all individuals. Typically, travel models that include bicycling as an alternative assume that it is available for all trips within a certain travel distance threshold, without regard to traveler attributes. Analysis of the 2011 Oregon Household Activity Survey data, used to estimate the model, revealed that just 29 percent of sampled residents of the Portland region, age 16 and over, reported owning and using a bicycle on a regular basis, emphasizing the need to consider individual attributes when modeling whether bicycling is a real mode alternative. While there are many studies that examine the effects of person attributes and urban form on the frequency of bicycling for various trip purposes, they do not address the more fundamental question of whether bicycling is considered a viable mode option for persons who are not observed bicycling.

Using a binary logit specification, we found that gender, age, commuting distances to work and school, job schedule flexibility, workplace parking costs, household income and car ownership, and the presence of children were significant predictors of whether a person of driving age said they own and use a bicycle on a regular basis. We also found significant effects of home-TAZ intersection density and accessibility to arts and entertainment employment, a proxy for indoor recreation opportunities, suggestive of an urban lifestyle. Further, the number of other adult household members who reported owning and using a bicycle on a regular basis had the most significant impact on model fit, highlighting that bicyclists tend to live with other bicyclists. In model application, this self-referential aspect of the model is handled by iterating over household members, a method that converges rapidly. Finally, we will show how conditioning mode choice sets on individual bicycle ownership affects estimated tour mode choice parameters and implied values of travel time.