How to better assess whether sustainability agreements are beneficial

How to better assess whether sustainability agreements are beneficial

What should competition authorities do when an agreement between competitors would both promote sustainability and restrict competition? In this article, Nadine Watson explains that – even within the consumer welfare framework – authorities can already support more sustainability agreements. Drawing on lessons from environmental economics, she demonstrates that the apparent tension between consumers’ interests and the environment is often a failure in analysts’ methodology, not in consumers’ values.


Increasingly, companies want to offer sustainable products and services. That can require cooperation between competitors, for instance, where they need to agree to combine their knowledge or assets to provide more sustainable services (“sustainability agreements”). Competition authorities also want to promote sustainability and Commissioner Vestager has urged competition enforcers to play their part, especially in the fight against climate change.[1]

However, it is not clear what role enforcers should play. Obviously, they must verify whether a proposed sustainability agreement would genuinely benefit the environment – enforcers cannot allow competitors to smuggle collusion behind “green-washed” façades. They must also verify whether these restrict competition and, if so, assess the trade-off of the expected benefits against expected anticompetitive effects.

When evaluating this trade-off authorities face an apparent constraint. Although sustainability promotes society’s welfare, authorities must restrict their attention to the benefit of the consumers of unsustainable products when assessing whether on balance sustainable agreements are anticompetitive. Despite the benefits to society, authorities cannot support sustainable agreements that do not generate sufficient benefits to consumers of unsustainable products to overcome the restriction to competition. This has sparked a debate on whether competition policy needs to change, to consider benefits that fall outside the framework of consumer welfare, when assessing sustainability agreements.[2]

In this article, we explain that, even within the current consumer welfare standard framework, authorities can support more sustainability agreements than is currently assumed, if they choose the right analytic tools. Often, the problem is not that consumers of unsustainable products do not value sustainability; the problem is that analyses underestimate how much they value it. Drawing on lessons from environmental economics, we find that the tension between consumers’ interests and the benefits to society can be a failure in analysts’ methodology, not in consumers’ priorities.[3]

How does promoting sustainability affect consumers’ welfare?

The goals of promoting sustainability and protecting consumers are not necessarily incompatible.

Some sustainability agreements will pose no threat to competition at all. To help identify these cases, the Dutch competition authority (the ACM) has already published guidelines illustrating such circumstances. For instance, it explains that waste collection companies with a combined market share of 30% could jointly optimize their routes to reduce costs, traffic congestion, and carbon emissions.[4]  The environment would benefit, and the market would remain competitive. Clarifying where such cases exist is useful. It tells companies where they can safely co-operate without fear of penalties.

Additionally, academics have argued that (in some circumstances) protecting effective competition is the best way to promote sustainability.[5] Where consumers desire sustainable products, companies should compete to attract those environmentally conscious consumers. The important point here is that improving sustainability can benefit consumers. The impact that a product has on the environment is an attribute of that product; it is part of the product’s quality that consumers might value.

However, there are circumstances when protecting competition and promoting sustainability would be in conflict. Competitors could agree to make their products and services more sustainable, benefiting the environment, society, and (potentially) consumers. But some of those agreements would increase production costs and/or exclude cheaper (polluting) alternatives, reducing choice and imposing higher prices on consumers.

Of course, competition law does not impose an absolute prohibition on mergers and agreements that constrain competition, even if they increase consumer prices. For instance, EU law permits restrictive agreements if they meet four cumulative conditions:

a. the agreement must contribute to improving the production or distribution of goods or contribute to promoting technical or economic progress;
b. consumers must receive a fair share of the resulting benefits;
c. the restrictions must be indispensable to the attainment of these objectives; and finally
d. the agreement must not afford the parties the possibility of eliminating competition in respect of a substantial part of the products in question.[6]

For sustainability agreements, the most challenging of these conditions is that “consumers must receive a fair share of the resulting benefits”. Authorities apply this requirement such that an agreement that restricts competition must provide positive effects to consumers (such as improvements in the quality of a good or service) that at least compensate them for price increases. So, in the case of a sustainability agreement, consumers would have to value the incremental environmental benefits more than they dislike the increase in prices.

Analysing the economic benefits of sustainability agreements in this way might strike us as unreasonably restrictive. There are two concerns that are worth considering separately:

a. How welfare is measured: Authorities measure consumer welfare by analysing how much consumers are willing to pay for an improvement in quality, as that reveals how much the proposed improvement is worth to them. The problem with sustainability is that people seem unwilling to pay as much to protect the planet as they should do, for instance, given the financial impact that studies suggest climate change will have on their lives.[7]
b. Whose welfare is measured: Authorities consider consumers’ welfare, but promoting sustainability is not intended to (only) benefit consumers; it benefits wider society and future generations. If consumers are unwilling to pay sufficiently more for sustainable alternatives, should competition policy really protect their right to buy polluting products?

Importantly, we don’t need to solve both issues to make progress; by ensuring that methodologies are used that accurately capture consumers’ willingness to pay for environmental benefits, more sustainability agreements would be recognised as beneficial.

Choosing the right methods to measure what consumers value

Authorities use various methodologies to measure how much consumers are willing to pay for products, or specific attributes of those products. We can group them into two categories.

a. Revealed preference methods. With these approaches, analysts observe consumers’ behaviour to infer how much they value the various attributes of a product. For instance, by analysing variations in purchase data for a product, analysts can infer how different attributes affect how much consumers are willing to pay for it.
b. Stated preference methods. With these approaches analysts ask consumers directly, for instance using surveys, how much they are willing to pay for products with various attributes.

Methodologies that reveal a lack of information, not a lack of value

Revealed preference methods are commonly applied in competition assessments and in most cases they should be reliable. For instance, consumers of coffee might not consciously know how much more they are willing to pay for better tasting coffee, or coffee that takes less time to make in the morning, but we can analyse data on those attributes, sales, and prices to reveal that information. Where data is available, these approaches have the advantage of being relatively quick and cheap to perform and they are grounded in consumers’ actual behaviour.

However, they also require that consumers experience the attributes they value. For instance, a coffee consumer knows what various brands taste like and how long each one takes to brew in the morning – even if they are unaware of those qualities initially, they can acquire that knowledge through consumption. We can reasonably assume that coffee consumers are sufficiently informed about the attributes of coffee that their purchases reveal their preferences accurately.

We can’t make the same assumption for sustainability. That is because consumers don’t need to experience environmental benefits in order to value them. Environmental Economists describe this phenomenon by distinguishing between “use value” and “non-use value” (or “passive value”). For instance, if someone valued preserving the Alaskan wilderness because they wished to experience hiking there, that environmental benefit would provide them with “use value”. However, environmental benefits commonly provide people with ‘passive value’. For example, the same person might also value the Alaskan wilderness for its own sake – simply because they value knowing that it exists or that it will be bequeathed to future generations. Passive value is entirely independent of a person’s desire to experience that benefit. Someone might passively value protecting the Alaskan wilderness despite not having any desire to experience it, for instance, if their fear of cold, hiking, and bears would make any trip there an unpleasant one.

This distinction matters for competition authorities, because they can only consider the value that methodologies are able to measure. Over fifty years ago, John Krutilla’s influential paper noted that consumers’ purchases would rarely reveal their passive-use value.[8] This is because environmental benefits are rarely (if ever) traded directly, and we are often unaware what impact our consumption has on the environment, so analysing our purchases will not reveal (even indirectly) our views on that impact.

For instance, imagine an environmentally conscious coffee consumer has no idea which brands use packaging that releases pollutants into the oceans, nor which brands have the highest carbon footprint. No amount of consuming coffee will provide that information to them. They will only be able to distinguish between sustainable brands and polluting brands if that information is provided to them. If it isn’t, analysing their purchases will not reveal how much they value brands that benefit the environment; it will reveal that they have limited information about the impact their coffee consumption has on the environment. Any meaningful analysis of consumers’ decisions needs consumers to be informed about how their choices affect their interests.

Measuring passive value

Environmental economists have used stated preference methods, such as choice modelling, to accurately measure how much people are willing to pay to protect the environment.

In 1989, for instance, the Exxon Valdez ran aground off the Alaskan coast, spilling 37,000 tonnes of oil. To establish how much Exxon should pay in damages, the state of Alaska commissioned a study to estimate how much the American public would be willing to pay to avoid similar disaster. Few Americans experience pollution in the Arctic directly, but nonetheless, many of those surveyed stated that they valued protecting that wilderness and would be prepared to pay a federal tax to protect it.

How the survey is conducted matters

Just because revealed preference methodologies understate passive value, does not mean that all survey-based methods will necessarily measure it accurately. The standard challenges faced when conducting surveys to measure willingness to pay and how they can be overcome can be illustrated in the case of the Exxon Valdez oil spill.

a. Respondents struggle to value unfamiliar attributes: without additional information, survey respondents are no more likely than consumers to know how much they value the environment, or how their actions would affect it. However, surveys provide an opportunity to inform respondents and to ensure their responses reflect considered views.[9] In the case of Exxon Valdez, researchers provided information from scientists and engineers to explain what impact another oil spill would have on wildlife and the shoreline (two aspects respondents were known to care about). They also provided visual aids to explain how ‘escort ships’ would prevent damage – which involved immediately deploying a ‘sea fence’ to contain the oil spill and use a ‘floating hose’ to recover it. The information was delivered so that it could be understood by respondents without a high school diploma.
b. Respondents will provide different answers, depending on how a question is framed: This phenomenon has prompted some critics to assert that surveys in general are unreliable. In fact, it demonstrates that survey design matters – there are good surveys and poor surveys.[10] For instance, respondents must believe that their answer will be consequential. If they believe that their answers have no consequences, then any potential answer will be as good as any other.[11] In the case of Exxon Valdez, the survey needed to provide credible choices, with well-defined benefits and costs, that respondents believed might be enacted. To ensure that, the survey described the proposed escort ship programme, its scope – i.e. that it would only prevent spills in Prince William Sound, on the south coast of Alaska – and that it would be funded by a one-off tax in addition to regular federal taxes.
c. Respondents exhibit biases (“behavioural effects”): Survey responses are not always in line with respondents’ behaviour. For instance, respondents often say they are willing to contribute more to the public good than they actually do. However, this would not undermine the survey response as it is exactly how economic theory predicts that people will behave when they can free ride on the contributions of others—they will contribute less than they are actually willing to. Similarly, if respondents are willing to contribute to protect the environment but lack faith that others will make the same sacrifice, then they will be less likely to contribute. The benefit of surveys is that they allow these attitudes and biases to be identified and accounted for. For instance, in the case of Exxon Valdez, respondents were asked if they believed the escort ship program would reduce the damage from a large oil spill by “a great deal”, “a moderate amount”, “a little”, or “not at all”. Unsurprisingly, respondents were willing to pay less for escort ships when they believed the proposal would be ineffective. Those respondents did value protecting Prince William Sound; they were just pessimistic that it would be protected.

The Exxon Valdez case successfully overcame these concerns. The U.S. government convened a panel, led by Nobel Laureates Kenneth Arrow and Robert Solow, to assess the methodology used to measure passive use. It concluded that the stated preference methods employed produced reliable estimates.[12] Based on the study’s results, Exxon agreed to pay approximately USD 3 billion.

The significance for competition cases is that, by conducting this type of analysis, already used in the assessment of mergers,[13] competition authorities could already do more to support sustainability agreements under the current framework.

Whose welfare should we analyse?

At the heart of the debate about sustainability agreements is whose welfare we should be concerned about: consumers’ or society’s. This question is closely related to how the relevant market is defined. It is worth considering three groups separately:

  • consumers (i.e., those in the relevant market of the unsustainable product the sustainability agreement will modify);
  • potential consumers (i.e., those in a prospective relevant market including the unsustainable and sustainable version of the product); and
  • non-consumers (i.e., those outside the relevant market of the unsustainable and sustainable version of the product).
The importance of defining ‘consumers’

It should be uncontroversial that authorities analyse consumers’ welfare and rely on analysis that does so accurately. However, it may not be clear which consumers one should analyse because relevant market precedents may not be applicable.

For some cases – such as considering the carbon emissions of different brands of coffee – the outcome of the analysis may not differ if one uses a very narrow market definition (e.g. instant coffee) or a broader one (e.g. all hot beverages). In other contexts, market definition may affect the outcome of the analysis a great deal. For instance, if we only consider the preferences of consumers of diesel cars we might find they value protecting the environment much less than if we analysed the preferences of consumers of cars in general.

The important point is that one needs a robust market definition, and that depends on which products consumers consider to be close substitutes. That question can only be answered if consumers are fully informed about rival products’ prices and characteristics, including the impact each one has on the environment. As a consequence, the question of market definition cannot be addressed separately from the survey of consumers’ preferences we discussed above: it should be informed by it.

Analysing the interests of “potential” consumers

Often, it is reasonable to analyse customers in the near-past to consider the interests of consumers in the near-future. However, there are two problems with this assumption in the context of sustainability. Firstly, future consumers (even in the near-future) may be willing to pay a lot more than current consumers, either as the effects of climate change become more apparent, or because they will (on average) have to live with its consequences for longer.

Secondly, when the quality of a product changes, the community of relevant consumers should expand, so that it includes consumers who would not buy the unsustainable version of a product, but would buy the sustainable version of that product.[14] In essence, these ‘potential’ consumers are willing to pay for the product only if it is sustainable.

The interests of potential consumers are a relevant consideration, but they will not be revealed by analysing customers’ historical purchases; they can only be established by stated preference methods. The significance of potential consumers will vary on a case-by-case basis. Where consumers are generally unaware of the impact a particular product has on the environment, there may be few customers that are aware of the negative impact a product has and deterred by that impact. However, for many products, it is possible that consumers’ awareness and preferences both vary, such that some of them lack information about their choices, while others are deterred by the information they have.

Including the interests of non-consumers would require substantial change

The most difficult – and substantial – debate concerns the interests of non-consumers. In normal circumstances, we don’t need to consider non-consumers’ interests. People don’t value the taste of someone else’s coffee; it doesn’t affect them. In contrast, people do value the impact that someone else’s coffee consumption has on the environment. It does affect them; they suffer the costs of pollution.

The crux issue is that consumption imposes externalities on society. The ‘benefits’ of sustainability might be described better as the ‘reduction in harms’ caused by consumption. Sustainable production methods don’t necessarily increase costs. Unsustainable production methods are already costly, but the costs of pollution are not borne by consumers, they are borne by society without compensation or consent.

The best way to deal with externalities is through regulation – such as imposing a carbon tax, equal to the harm that carbon imposes on others. Given the difficulty of doing that, it can be tempting to see competition enforcement as a complementary tool to address externalities. For instance, the requirement that “consumers receive a fair share of the benefits” could take on a different meaning. If production costs reflected the cost of pollution, consumers would already pay higher prices. In this light, requiring that consumers are fully compensated for reducing the harms they push onto society does not seem fair; it seems blatantly unfair.

However, the repercussions of adopting such an approach are both complicated and uncertain. To avoid unintended and negative consequences, any proposed change requires careful consideration – it is not good enough to the be right in principle, it must also be effective in practice. Even then, such an approach would require changes to the current framework, adding further time and complexity.

In the meantime, we can make progress under the current framework. Simply by insisting on methods that measure consumers’ genuine preferences accurately, we can better protect their interests, and those of the planet.

Read all the articles from the Analysis

[1] Vestager, Margrethe, “The Green Deal and competition policy”, 22 September 2020. Available at (downloaded 15 March 2022):

[2] Holmes, S., D. Middelschulte and M. Snoep (eds.), Competition Law, Climate Change & Environmental Sustainability, Concurrences, Paris,

[3] Recently, the same argument has been made by Inderst and Thomas; namely, that “a consumer’s choices could fail to express her own preference, e.g., due to a lack of information”. Inderst, Roman and Thomas, Stefan, Measuring Consumer Sustainability Benefits (September 1, 2021). Available at SSRN: or

[4] ACM (2021), Draft guidelines ‘Sustainability Agreements, second draft, p. 16.

[5] Schinkel, M. P. and L. Treuten (2021), “Green Antitrust: Friendly Fire in the Fight against Climate Change”, Amsterdam Center for Law & Economics Working Paper No. 2020-07. Available at

[6] European Commission, Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreements, OJ 2011/C 11/4.

[7] Dolmans, Maurits (2020), Sustainable Competition Policy, CLPD Competition Law and Policy Debate Vol 5, Issue 4 and Vol 6 issue 1 March 2020; Holmes, S. (2020), “Climate change, sustainability, and competition law”, Journal of Antitrust Enforcement, Volume 8, Issue 2, July 2020, Pages 354–405.

[8] Krutilla, J. (1967), “Conservation Reconsidered”, The American Economic Review, Vol. 57/4, pp. 777-786.

[9] Carson, R. (2012), “Contingent Valuation: A Practical Alternative when Prices Aren’t Available”, Journal of Economic Perspectives, Vol. 26/4, pp. 27-42.

[10] Carson, R. and T. Groves (2007), “Incentive and Informational Properties of Preference questions”, Environmental and Resource Economics, Vol. 37/1, pp. 181-210.

[11] Carson and Czajkowski (2014), “The discrete choice experiment approach to environmental contingent valuation”, in Handbook of Choice Modelling (eds. Stephane Hess and Andrew Daly), Edward Elgar Publishing.

[12] Arrow, K. and E. al. (1993), Report of the NOAA Panel on Contingent Valuation, Federal Register.

[13]  Case M.7758-Hutichinson 3G Italy/Wind/JV.

[14] Dolmans, M. (2020), “Sustainability Competition Policy and the ‘Polluter Pays’ Principle”, in Competition Law, Climate Change and Environmental Sustainable, Concurrences.