Income explains much of Scotland’s EV charging divide

A new study in Transportation Research Part A applies causal inference to nationally representative Scottish household data. The authors estimate that income accounts for much of the apparent relationship between off-street parking and electric vehicle adoption. In their model, providing off-street parking raises EV ownership from 3.3% to 5.6%, while moving a household from the lower to the higher income stratum reduces market non-participation by 23.1 percentage points. For developers working with survey data, the study shows why a strong predictor may still be a weak policy target.

Why driveways absorb the income signal

Off-street parking makes overnight charging cheaper and more convenient, so households with driveways tend to own EVs at higher rates. Income complicates that association. Higher-income households are more likely to have private parking, afford an EV, qualify for finance, and absorb the risk of adopting a relatively new technology. A regression that incompletely accounts for those factors can assign parking an effect partly generated by wealth.

The study represents this structure with a directed acyclic graph, or DAG. Income Z points to both parking access X and EV adoption Y, creating a back-door path between parking and ownership. Conditioning on a sufficient set of common causes blocks that path under the graph’s assumptions, allowing the researchers to estimate the distribution associated with an intervention: P(Y | do(X)).

How survey columns become interventions

The researchers combine automated causal discovery with domain review. Their MIIC algorithm builds an undirected graph from information-theoretic dependence tests, then uses interaction information among variable triples to orient eligible edges. Under its orientation rules, positive interaction information supports a chain or common-cause pattern, while negative interaction information suggests a collider, where two causes point into the same effect.