I estimate the relationship between income, frequency of exposure to air pollution and the composition of air pollution (’cocktail days’) using French municipal data over the period 2012-2018. I propose a new method for assessing exposure to air pollution cocktails, based on two new indicators that take into account simultaneous and multiple exposure to several air pollutants. I find contrasting results between rural and urban areas. The most affluent urban municipalities have on average a lower number of pollution days compared to the poorest urban municipalities. However, the pollution days are composed of an equal proportion of cocktail days between the poorest and the most affluent urban municipalities. In rural areas, on the other hand, wealthier municipalities have a higher number of pollution days, composed of a higher proportion of cocktail days than poorer municipalities. I also show that pollution levels are higher overall in urban areas, as well as the proportion of cocktail days. In addition, the exposure gap is greater in urban areas.

This paper assesses the short-term impact of air pollution on human capital using comprehensive data on the results of the two main national examinations in France, the “Brevet” (9th grade) and the “Baccalauréat” (12th grade), for cohorts from 2010 to 2017. Taking advantage of random daily variations in air pollution between exam days, we show that a concentration of fine particles (PM2.5) above the WHO threshold significantly decreases the exam scores of the poorest pupils and those experiencing initial difficulties at school. This result is robust when the instrumental variables method is used, exploiting daily variations in wind direction. Taking the Brevet exam on a day when PM2.5 levels exceed WHO recommendations results in a 5% reduction in the standard deviation of the exam mark for the poorest pupils. These pupils also have the greatest difficulties at school, further exacerbating inequalities in education.