Do increases in local air quality lead to higher labor productivity?

Air pollution, in particular PM_2.5, might reduce short-term worker productivity via increases in blood pressure, headaches, respiratory irritation and irritation to the eyes, ears, nose and throat. PM_2.5 is approximately 1/30 the diameter of a human hair, and can enter buildings easily, with penetration rates from 70–100 percent.

Most of the previous literature has looked at the impacts of long-term exposure to pollution on health.  However, the detailed data required to isolate the more direct impact of pollution on contemporaneous labor productivity, separately from overall impacts on health, poses a significant barrier to research. Graff Zivin and Neidell (2012) use worker-level daily productivity data for a large California farm matched with local ozone air quality monitoring data. They find “that ozone levels well below federal air quality standards have a significant impact on productivity: a 10 parts per billion (ppb) decrease in ozone concentrations increases worker productivity by 5.5 percent."  Chang et al (2014) examine daily productivity of indoor workers at a pear-packing factory, and find that an increase in outdoor levels of PM_2.5, leads to a significant decrease in productivity. Li, Liu, and Salvo (2015) use data on workers from a manufacturing plant in the Hebei province of China, and find large reductions in indoor worker productivity (on the order of 15%) associated with the first 200 µg/m3 increase in concentrations of fine particles (PM_2.5).

We estimate the impact of changes in local air quality on labor productivity, by combining hourly data on ambient concentration levels of specific air pollutants – particulate matter (PM_2.5), ozone, sulfur dioxide, nitrous oxides – with high frequency data on indoor labor productivity at supermarkets in a western U.S. metropolitan area, using a dataset from Mas and Moretti (2009). Our panel consists of all transactions performed by cashiers from six stores of a national supermarket chain for two years, for dates between 2003 and 2006. We have scanner-level data on the number of items scanned and the length of each transaction, measured in seconds, which allows us to obtain a precise measure of productivity for each cashier. We measure hourly concentration levels of O₃, PM_2.5, SO₂, and NOx with air quality monitors located near the supermarkets. Our empirical strategy exploits the high-frequency fluctuations across time and space in ambient concentrations of these pollutants.