8. Discussion

After performing the empirical estimations, it can be concluded that neither basic nor advanced low emission zones have an effect on average PM10 concentrations. This applies not only for the group of treated stations as a whole but also when analyzing the effect on stations with respect to different characteristics, as it is done in the models introducing heterogeneity to the treatment group. While Demonstrativethese concentrations are not problematic with respect to the EU limit value of 40µg/m³ (compare table 3), they are still on average a lot higher than the health-based target value of 20µg/m³ per year suggested by the World Health Organization (2006).[/annotax]

However , when it comes to reducing the number of days that record an average 24 hour concentration above 50µg/m³, the introduction of a basic LEZ seems to reduce the exceedance days by an average of 5 days per year. This value coincides with the reduction that was expected by the Federal Environmental Agency (2012a) in an attempt to model the effects of LEZs. But does that mean that all cities struggling with that limit value should introduce an LEZ? Or all cities in general, as the WHO recommendations of zero days in exceedance of 50µg/m³ would suggest? As the analysis of the heterogeneity within the treatment group shows, this should not be concluded. The policy seems to work only for cities that basically constantly violate the EU standard or for large cities with more than 500,000 inhabitants (the groups overlap by 60%). Yet , for those cities, the policy seems to be quite effective with a reduction by 8-9 days per year. Considering the whole picture, the results question the wide-spread introduction of LEZs that could be observed in the last three years; it seems insufficient to rely only on LEZs when trying to improve air quality. Basic LEZs, by excluding the dirtiest vehicles, succeed under specific conditions to reduce “peak” concentration days where pollution is excessively high, but fail to reduce average concentrations. It should also be noted that even if a reduction of 8-9 days per year is achieved in highly polluted cities, this does not automatically ensure compliance with EU law: the stations in the sample recorded up to 191 violations of the 24h limit per year, as a closer look at the distribution of the number of days in violation of the 24 hour limit shows (see table A2 in the appendix). Thus , low emission zones are no quick-fix for compliance, but can under certain circumstances serve as one part of a solution.

Against the expectations of policy makers, the introduction of the stricter standards of an advanced LEZ does not lead to additional significant improvements of air quality in comparison to introducing only a basic low emission zone. The empirical results show  neither a reduction in average air pollution, nor non-compliance days with the 24 hour limit, while the Federal Environmental Agency (2012a) expected the introduction of a LEZ step 3 to reduce yearly average concentrations by 10-12% and non-compliance days by around 20 per year. These expectations cannot be confirmed by the results of this study (note, however , that the indicator used for an “advanced” LEZ summarized LEZs step 2 and 3). The results should be handled with care though and be re-examined when more data on advanced LEZs becomes available, since the sample size was small and the observation period rather short. However , the preliminary results can give some implications that should be kept in mind when designing future policies: the fact that advanced LEZs do not improve the pollution situation further compared to basic LEZs could be a hint that potential gains from a stricter technology standard are offset by other effects. A rebound effect, such as people driving more or longer due to a cleaner and probably more fuel-efficient car, could be considered one explanation. Even if it is not possible to draw definite conclusions about such an effect from this study, the results raise questions on the effectiveness of technology-based driving bans in comparison to policies that aim at actually reducing driving activity. However , as the failure of Mexico City’s driving ban shows, such policies have to be carefully designed as well (Davis, 2008). The cities of Stockholm and London claim positive experiences in reducing particle pollution with road pricing schemes, which can provide incentives to drive less, as can well-designed public transportation schemes (Eliasson et al., 2009, Atkinson et al., 2009).

Another issue that could not be taken into account in this study is the one of compliance. It is not possible to assess if the rather small effects result from the actual policy design or from non-compliance with the policy, since it is arguably difficult to control the adherence of drivers with the regulation. Another question that cannot be addressed in this study but could be important when discussing air pollution from traffic is the appropriateness of the EURO exhaust standards, on which the LEZ design builds upon. Another interesting extension of the study would be to compare the effects of German LEZs to LEZs implemented in other European countries. Currently, LEZs operate in eleven countries, differing to various degrees in design and strictness (Low Emission Zones in Europe Network, 2012). Exploiting this variation could provide interesting insights into the functioning and effectiveness as driving bans, but has to be left to further research.

With regard to the validity of the estimates, the biggest threat arises from unobserved, time-varying factors or political preferences that are correlated with both treatment assignment and outcome. The graphical analysis of the data showed that the treatment and the control group are not driven by differing underlying trends, remediating this concern. Through the large control group that includes non-treated stations in all states that host LEZs , state-specific policy preferences can be controlled for. Yet , it remains an open question why LEZs are especially popular in some federal states but not in others, and could be an interesting aspect for further research. The spatially imbalanced distribution of LEZs also requires an estimation strategy that can control for possible omitted factors arising from this aspect, which in this study was handled by region-specific fixed effects. The sensitivity analysis done in part 7.4 showed that the results are insensitive to including variables that capture changes in local economic conditions. To control for meteorological conditions would be another important aspect, but requires the availability of spatially highly disaggregated meteorological data. Due to time constraints, such data collection has to be left for further research.

9. Conclusion

In this paper, the first systematic evaluation of Germany’s low emission zones as a policy to reduce PM10 pollution from traffic sources has been conducted. Using a difference-in-difference strategy on station-level PM10 panel data for 2004–2010, it was tested both for average and heterogeneous effects of low emission zones on two different outcome variables. The results show that basic low emission zones do not have an effect on average PM10 concentrations. Yet , the policy reduces the number of days with excess pollution, as expressed by the 24 hour limit value of the European Union, in large and highly polluted cities by 8-9 days per year. These positive effects seem to be limited to the zones itself, as no spillover effects to adjacent measuring points could be detected. This finding also limits possible health benefits of LEZs, but at the same time rejects the concern that LEZs cause additional pollution around the zones, for example by creating longer driveways. Against political expectations, the stricter technology standards required by advanced LEZs did not show any significant effect in the data at hand. This is especially interesting because most of the cities currently having a basic LEZ in place plan to tighten their standards and introduce a LEZ step 2 or 3 (summarized as “advanced” zones in this study) in 2013 or already have done so in 201216. Considering the results from this study, it is questionable if relying on LEZs will lead to the desired effect and to compliance with European law; it seems rather likely that additional measures are needed to reduce pollution to the required levels. An advice to policy makers would therefore be to carefully evaluate the nature of the pollution problem in the respective city and consider alternatives to low emission zones, instead of relying on LEZs as the ultimate tool to reduce PM10 pollution.

The results of this study also contribute to the research on transportation policies by adding empirical findings on a technology-based driving ban. Though there is no definite evidence of a rebound effect, the missing effects of advanced LEZs on pollution levels raise questions with regard to behavioral responses that might offset the technological gains. Further research is needed to determine the factors that actually influence driving behavior of people and their reactions to policy, but the findings of this study challenge the suitability of command-and-control policies. If even strict technology standards prove to fail in reducing air pollution, there might be a need for policies that ultimately aim at reducing driving activity to reach air quality targets.

Footnotes

16 See Federal Environmental Agency (2012d) for a list of planned and current LEZs.