Zero‑Vehicle, but Not All the Same: What Are Their Real Travel Habits?

When we think of families without a car, we often picture people walking or taking the bus for everything. The newly published study breaks this stereotype: Zero‑Vehicle Households (ZVH) actually consist of three distinct groups. The first, called “shared mobility errand workers” (36 % of cases), mainly uses public transport and ride‑hailing apps for work and daily chores. The second, the “car‑based shoppers” (30 %), don’t own a car but have informal access – friends, family, or car‑sharing – for longer trips like outings or big shopping, mixing them with walking or cycling. The third group, the “active shoppers” (34 %), prefers walking or biking for local grocery runs.

These differences were uncovered using a statistical method called Latent Class Cluster Analysis, which examined travel modes, trip purposes, and socio‑economic variables. The result is a more realistic picture that helps urban planners design tailored solutions: more ride‑hailing pick‑up points, shared parking for those who use others’ cars, and safer bike lanes for cyclists. In practice, understanding ZVH nuances means building cities that are more inclusive and less dependent on private cars.

The analysis draws on a national sample from the 2022 National Household Travel Survey, which records over 30,000 daily trips. Latent Class Cluster Analysis (LCCA) is a statistical modeling technique that identifies latent sub‑groups within a heterogeneous population by combining observed variables (transport mode, trip purpose) with covariates (age, income, built‑environment density). The model was weighted to correct sampling bias, ensuring the classes reflect the real population.

The three emerging classes display distinct mobility patterns: the “shared mobility errand workers” class has a high probability of using ride‑hailing services (Uber, Lyft) and public transport for work and errands, averaging 2.3 trips per day. The “car‑based shoppers” class relies on informal vehicle access, with an average of 1.8 long‑distance trips (>10 km) per week, often for recreational activities or bulk shopping. Finally, the “active shoppers” class predominates short trips (<2 km) on foot or bike, strongly correlated with high‑density neighborhoods and proximity stores.

These findings have implications for travel demand modeling: traffic‑reduction policies must consider not only the presence or absence of private cars but also the availability of shared mobility services and the built environment configuration. Integrating LCCA data into forecasting models can improve traffic simulation accuracy and support more targeted decisions for pedestrian, cycling, and micro‑mobility infrastructure planning.