I combine measures of neighbourhood characteristics with high-resolution remote-sensing data to provide the first national-scale study of cross-sectional and longitudinal inequality in exposure to fine particulate matter (PM 2.5) in France. Descriptive evidence indicates that, at the national level, there is a U-shaped relationship between income and PM 2.5 exposure, which is not reflected within urban areas. Fixed-effect models confirm that on average, higher neighbourhood income is associated with lower exposure. Longitudinal inequality measures suggest that recent air quality improvements accrued predominantly to areas that had a lower initial exposure, and intermediate income. I then exploit a change in air quality schemes at the level of urban areas in an event-study framework, so as to shed light on potentially unequal benefits from the induced reduction in exposure. I find that initially lower-income areas received smaller benefits from the policy change, and quantile regression estimates suggest that exposure decreased more in less polluted areas. As some results are sensitive to formally accounting for spatial autocorrelation, this study also underlines the need to pay specific attention to this issue when measuring environmental inequality.