The interplay between residential location and cycling choice: the case of two metropolitan areas in Sardinia, Italy

The current paper aims to enrich the current understanding of the link between the choice of residential location, the propensity to cycle to work and the propensity to cycle for non-commuting purposes. To highlight the relationship among these choice dimensions we used a composite econometric model that allows for the joint modelling of multiple outcomes. Residential location and cycling propensities are modelled as a function of socio-demographic and level-of-service variables. The inclusion of common error terms allows us to control for self-selection and unobserved effects that can simultaneously influence the underlying propensities. The data for this study is drawn from a survey conducted in the metropolitan areas of Cagliari and Sassari (Sardinia, Italy) in 2016 among a sample of local employees. The sample comprises 2,128 observations. Our results indicate that a significant portion of unobserved variance between the residential location choice and the propensity to cycle for non-commuting reasons exists, suggesting the presence of a self-selection effect.