When individuals impose costs on others, without having any economic incentive to take those costs into account, we are in the presence of a negative externality. As explained by the economic theory, externalities may lead to individual decisions that are not optimal from the point of view of society as a whole. Without any intervention, the market will tend to overproduce negative externalities.
Air pollution is often presented as a classical example of a negative externality. In recent years, we have been confronted by the media with some news pointing out pollution problems in central Lisbon and claiming urgency towards action. Avenida da Liberdade, a key point for those who enter or leave the capital by car, is today one of the most polluted avenues in Europe. Several times, the legal admissible levels of air pollution were exceeded. In this context, not respecting the EU Directive of Air Quality might have severe consequences, especially when Portugal is under external assistance and so financially constrained. Would we like to pay a fine of 1.9 Million Euros plus 630 Euros per day until compliance with the EU Directive?
Furthermore, the problem is even worse due to the daily traffic congestion during rush hours. As we know, urban traffic congestion is a modern example of a “tragedy of the commons”. If we consider the trips by car along the Avenue as a good, immediately, we will agree on its non-excludable nature. However, since there is free access, there will be rivalry in consumption, that is, the access creates scarcity of space and thus congestion.
How should this problem be addressed? As a science, Economics can offer its contribution. Indeed, from very simple models, we can extract powerful insights. The 2012 Prize for Shapley and Roth clearly demonstrates the growing acknowledgment for the use of abstract theory to ongoing efforts to find practical solutions to real-world problems.
We decided to use a simple model with marginal benefits and marginal costs. The model incorporates both pollution and congestion problems. We have assumed that there is demand for a certain good Q, which will represent the number of trips by car along the Avenue. The cost structure has two major components: (1) private costs, including costs related to car use such as fuel and maintenance, and also time costs derived from traffic congestion that drivers partially take into account; (2) external costs, which include all those costs that drivers impose on others and that they have no economic incentives to take into account (e.g. reduced air quality, increased level of noise and higher risk of certain diseases, as well as the external portion of congestion that is not internalized). The sum of private with external costs gives the social cost function.
We have assumed that until QA there is enough environmental assimilative capacity to absorb the emissions associated with the number of trips in the Avenue, and that congestion only starts after that level. In other words, below QA there are no external costs to society (no pollution and no congestion problems), but only the private costs to drivers, which we considered as constant. Then, after QA, the marginal private cost function (MCp) becomes increasing due to the portion of time costs internalized by the drivers. Additionally, after that level, the number of trips along the Avenue starts causing external damages to the society: congestion not internalized (cong) and air pollution (e). It is important to note that in our setting, we have assumed complete independence between the damages of congestion and the damages of pollution. In practice, it is plausible that there is a relationship between these two functions, which would further complicate the analysis of the problem. Here follows the graphical analysis (the Q-axis is normalized):
The optimal number of trips in the Avenue is given by Q*, where marginal social costs (MCs) equal marginal benefits (MB). At that point, from a social perspective, there are no economic incentives to increase or decrease the number of trips. Obviously, if doing nothing is a possibility, drivers will freely choose Q0, where they are minimizing their private costs.
From our simple setting, we will be able to analyze, from an economic perspective, two alternative policies to solve the pollution problem.
The first one, which is already implemented by municipal authorities, consists in changing traffic laws in a way that access is restricted. Due to restriction of access, the number of trips that starts causing congestion will be lower than QA and marginal private cost will become an increasing function after that level. However, restriction of access will change the social optimum. Since the government is trying to solve the pollution externality by restricting the access to the Avenue, in reality, it is creating a bigger problem by correcting one externality aggravating the other. Now, as traffic starts congesting at an earlier stage, the new social optimum will be lower than Q*.
An alternative policy would be the imposition of a congestion charge, covering a vaster area in the center of Lisbon (including the Avenue), similar to the one already implemented in London since 2003 (for further details: www.cclondon.com). Theoretically, that charge would be a Pigouvian tax, which is, by definition, a unitary tax equal to the difference between marginal social costs and marginal private costs at the optimum. While raising revenue through the implementation of this tax, authorities force drivers to internalize the damages that were previously not taken into account. The net losses associated with higher private costs due to traffic congestion are now a net transfer of welfare from drivers to the government.
Both policies attain the objective of reducing the number of trips. However, restricting access to the Avenue leads to an inefficient outcome, since those losses incurred by drivers are not appropriated by the government. On the other hand, a Pigouvian tax will lead to an efficient outcome as it maximizes net benefits from a social perspective, that is, we will reach Q* again. Additionally, this measure will not aggravate the existing externalities.
In conclusion, our example illustrates the distinction between an efficient policy and a feasible one. While it is efficient to apply a Pigouvian tax, it might not be practical, since it imposes a higher burden to drivers than restriction of access.