Epsilon Archive for Student Projects

Abstract

This study attempts to estimate the causal effect of Sweden's first passenger car low emission zone on street-level PM10 concentrations. This is done by looking at the miljözon klass 2 which was introduced on Hornsgatan in Stockholm in January 2020, using a difference-in-differences design with Malmö's Dalaplan as a control street. Weekly mean PM10 concentrations are compared between the treated and control streets using a data set spanning from 2015 to 2024. The preferred specification yields a DiD coefficient of 1.574 µg/m³ (SE = 0.994), which is statistically insignificant at all conventional significance levels. A placebo test using Sveavägen as an alternative control street produces a similarly insignificant result. The event study reveals considerable pre-treatment coefficient volatility, raising concerns about the validity of the parallel trends assumption and limiting the conclusions that can be drawn from the estimates. Given these identification challenges, the results cannot be interpreted as definitive evidence either for or against the policy's effectiveness. The absence of a detectable effect is, however, consistent with documented enforcement failures. Stockholm's own post-implementation evaluation found that at most five fines were issued in the first year of operation despite widespread violations. This suggests that weak enforcement may have substantially undermined the zone's potential air quality impact. Future research should incorporate multiple treated units and exploit variation in LEZ adoption across Swedish municipalities to enable more credibly identified causal inference.