Do households substitute mileage across their cars?

Understanding the driving behavior of multi-car households is crucial to evaluate the effectiveness of policies aimed at reducing fuel consumption in the passenger car fleet. In this paper, we investigate whether households adjust their vehicle miles traveled (VMT) among their cars in response to changes in fuel prices. We use data from the 2017 National Household Travel Survey (NHTS) on two-car households, where both cars are powered by gasoline, and estimate econometric models that distinguish cars based on their fuel economy. Our results show that an increase in per-mile driving cost significantly reduces the demand for driving the same car and increases demand for driving the other car, with little to no evidence of overall substitution at the daily or annual level. However, further investigation using artificial monthly travel diary data revealed a pronounced substitution towards the most fuel-efficient car only when there is a large and sustained increase in gas price.