Rarely Enjoyed? A Count Data Analysis of Transit Ridership in Germany's Public Transport

Focusing on adult members of German households, this paper investigates the determinants of public transit ridership with the aim of quantifying the effects of fuel prices, fares, person-level attributes, and characteristics of the transit system on transport counts over a five-day week. The reliance on individual data raises several conceptual and empirical issues, the most fundamental of which is the large proportion of null values in transit counts. To accommodate this feature of the data, we employ modeling procedures referred to as zero-inflated models (ZIMs), which order observations into two latent regimes defined by whether the individual never uses public transport. The model estimates reveal fuel prices to have a positive and substantial influence on transit ridership, though there is no evidence for a statistically significant impact of the fare. Methodologically, ZIMs are seen to have superior predictive accuracy over the classical count data models, and thus may serve as the method of choice when the aim is to predict trip frequency for modes that a large fraction of the population never uses.