Can Building Subway Systems Improve Air Quality? New Evidence from Multiple Cities and Machine Learning

Public investments in subway systems are often partly motivated by improving local air quality and greenhouse gas emissions. Recent studies have investigated the air quality effects of subway investments, reaching differing conclusions across cities and periods. To reconcile these findings, we examine the air quality effects of all 359 subway system expansions in China between 2013 and 2018. The machine learning (ML) method adopted in this paper removes the variation of the high-frequency and seasonal air quality and therefore substantially improves the consistency and precision of the estimates. Based on the ML method, we find that although, on average, subway system expansions did not improve air quality in the short term, there is evidence of air quality improvement in the long term. This helps reconcile the different findings of the studies with different bandwidths. We also find that cities with low incomes or high economic growth experienced statistically significant improvements in air quality, which helps explain the different findings in the literature for cities with different characteristics.