In a recent paper I have evaluated for the first time the causal effect of the London Congestion Charge on the level of pollution. To this end, a unique dataset was assembled. This consists of daily observations, concentrating on eight pollutants: CO, NO, NO2, NOX, O3, PM2.5, PM10, SO2. By using a regression discontinuity design in time series; with thresholds centered on the dates of the introduction of the charge and of the beginning and end of Western Expansion, a negligible and adverse impact of the charge is documented. When a spatially disaggregated model is estimated, it emerges that the road pricing scheme has induced a decrease in the concentration of NO, NO2 and NOX in the charged area and an increase in surrounding areas. A general deterioration of pollution concentration is found in the case of O3, PM2.5, PM10. These results are consistent with an overall increase in traveled kilometers, due to traffic diversion from the charged to the uncharged area. Furthermore, there is an unclear, possibly adverse, impact of increased speed on pollution.
In particular, a significant decrease in the concentration of: NO, NO2, O3, PM10, SO2 was found in the treated area; and a contemporary increase in the concentration of: O3, PM2.5, PM10 was discovered out of the charged area. This pattern in the estimates is consistent with the hypothesis that the introduction of the congestion charge has diverted traffic in space and shifted drivers from charged to uncharged routes and, eventually, vehicles. Traffic data show that the number of circulating vehicles in the area surrounding the treated area by 279,596 vehicles; 242,441 of which were cars.
A substantial increase in the number of motorbikes and of bikes has also been detected in the city center. In terms of kilometers traveled, a decrease by 45,000 kilometers × vehicles was estimated in the treated area, along with an increase by 39,000 kilometers × vehicles in the surrounding area.
Overall, in terms of pollution concentration of the whole city, results suggest that the congestion charge has a limited impact. This is possibly because of the spatial diversion of traffic. This result is further corroborated by similar evidence existing for Milan and calls for careful consideration of the spatial extent and of the cross elasticity of traffic flows when implementing road pricing.