- Fresh objections have recently been raised against BMW, Daimler, Volkswagen, Audi and Porsche by the European commission, related to the potential collusion surrounding the implementation of emissions reduction technologies
- Although consumers would have been harmed by the collusion, it did not overtly involve price fixing or restraint of output, which creates added complexity for the purposes of economic damages modelling
- In this memo, we propose two approaches for quantifying damages based on the valuation of the emission technology, which could be implemented in a preliminary way using publicly available datasets
In October 2017, the European Commission announced that it had carried out inspections on the largest German auto manufacturers (BMW, Daimler, Volkswagen, Audi and Porsche, known collectively as the “circle of five”), on suspicion that the auto manufacturers had colluded to slow the rollout of technologies related to petrol and diesel car emissions. These inspections prompted the commission to begin a new investigation a year later, in September 2018, into the longstanding technical meetings that took place between engineers and managers at these companies.
In early April of 2019, the European Commission sent a Statement of Objections to BMW, Daimler and Volkswagen related to the meetings.1 EU Commissioner Margrethe Vestager said in a press release that while cooperation between competitors to improve technology was entirely acceptable, “EU competition rules do not allow them to collude on exactly the opposite: not to improve their products, not to compete on quality. We are concerned that this is what happened in this case.”
In the statement of objections, the Commission has outlined concerns regarding two main technologies;2
- Selective catalytic reduction (‘SCR’) systems – which reduce harmful nitrogen oxides emissions of diesel passenger cars (2006 to 2014)
- ‘Otto’ particle filters (‘OPF’) – which reduce harmful particle emissions from the exhaust gases of petrol passenger cars with direct injection (2009 to 2014).
How did companies collude to delay implementation?
Economists recognize that intra-industry cooperation in research and development can have procompetitive effects.3 When engineers are able to exchange ideas and discuss results, it can accelerate the rate of technological innovation and allow the creation of efficiency-enhancing standardization measures.
However, this same cooperation can also blend into collusion. According to antitrust theory, information exchange and transparency allow companies to monitor each other’s adherence to an agreement, and to coordinate production and pricing strategies to extract supracompetitive profits at the expense of consumers.4
In both cases, the circle of five’s technical meetings could have served, explicitly or implicitly, as a venue for the companies to monitor each other’s progress in implementing these technologies. Such monitoring is crucial for the functioning of a cartel.
How were consumers harmed?
Restricting competition on innovation would violate EU competition rules, which prohibit cartel agreements to limit or control production, markets or technical development.5 However, a collusive cartel to delay the implementation of emissions standards presents a challenge for modelling the financial harm suffered by consumers.
It is true that in a counterfactual world, in which the delay had not occurred, consumers would have been marketed more efficient vehicles. But it is very unlikely that they would have paid less for these counterfactual cars, meaning that a naive calculation of harm (what consumers paid versus what they would have paid), could very easily result in a negative damage figure. Nevertheless, real harm did occur, in that consumers lost out on having more technologically advanced vehicles available to them.
What might a model of harm look like in practice?
Analysis of individual component value
One potential approach to resolving the apparent contradiction discussed above comes from the damages calculation methodology adopted in defective product cases in the US.6 In such cases, plaintiffs’ experts looked at the resale value of phones in the secondary market in order to determine the loss in value that results from a “defective” product. What a claimant in a hypothetical emissions case and a claimant in a defective product case have in common is that they have been harmed only in the impaired or reduced function of a specific component of the product they are consuming.
Estimating the value that consumers attach to a lower-emission vehicle could be derived using multivariable regression or option demand model with distinct variables for things like fuel efficiency, number of seats, horsepower, brand, class and transmission. Estimation of total price as a combination of these variables would give an indication of the incremental cash value that consumers place on each, including a coefficient for lower emissions.
Damages could therefore be calculated as the value that consumers would have realized, evaluated in terms of their own revealed preferences, but for the purported scheme to slow the EU’s emissions standards.
Analysis of resale value
Another, more straightforward and less data-intensive route to quantification of lost consumer value may run through resale figures in the secondary automobile market. A comparison of the resale value of cars with and without the SCR/OPF systems would give a direct indication of their cash value. The reasoning here, assuming a cartel was active in restraining the rollout of the technology, is as follows:
- In the counterfactual world, consumers would have paid the same or close to the same price for the more technologically efficient cars. The evidence for this is the very fact that the circle of five acted to slow the implementation; had auto manufacturers felt they could extract the costs of implementation from consumers in the form of commensurately higher prices, the slowing would have been counterproductive.
- Having bought their cars, even if not all consumers would have drawn utility value from the better emission technology, they all certainly would have been able to benefit from the better technology to the extent that these cars were worth more on the secondary market.
- Adjusting for factors like mileage and age, the additional value to consumers per vehicle can be computed as the difference in price between the older cars that lack the emissions technology and newer cars that have it.
- The total lost value to consumers can be calculated as the number of sold cars that would have employed the SCR/OPF technology multiplied by the lost value per vehicle.
The European Commission has not yet concluded that the circle of five colluded to hold the SCR and OPF technologies off the market. However, should such a decision emerge, plaintiff classes may find it challenging to present a plausible model of harm using conventional damage calculation techniques, given that the counterfactual world, paradoxically, may have had similar or even higher prices than the actual world.
The techniques discussed above, which approach damages in terms of a cash-denominated evaluation of lost consumer value, may represent a way around this problem. Fideres’ analysis and review of available sources suggests that a preliminary analyst might be possible using publicly available data.
5 This prohibition is laid out in Article 101(1)(b) of the Treaty on the Functioning of the European Union and Article 53(1)(b) of the EEA Agreement
6 See, for instance, In re: Nexus 6p Products Liability Litigation or Christina Grace et al. vs. Apple Inc.
Max joined Fideres in 2016. He has led the development and implementation of economic models for a range of high-profile cases in the US and the UK, contributing to litigation on a variety of topics. He has contributed economic work to cases involving FX rate manipulation and LIBOR suppression financial antitrust and financial benchmark manipulation), the abuse of dominance by major US hospital systems (healthcare antitrust), and supplier cartels in agricultural industries (agricultural antitrust). Before joining Fideres, Max worked at the national laboratory in Los Alamos, New Mexico, as part of a team designing neural networks for applications in machine learning. Max holds an MSc in Economic History from the London School of Economics.