Market nudges and autonomy

Abstract

Behavioural techniques or ‘nudges’ can be used for various purposes. In this paper, we shift the focus from government nudges to nudges used by for-profit market agents. We argue that potential worries about nudges circumventing the deliberative capacities or diminishing the control of targeted agents are greater when it comes to market nudges, given that these (1) are not constrained by the principles that regulate government nudges (mildness, sensitivity to people’s interests and public justifiability) and (2) are often ‘stacked’ – they come in great numbers that overwhelm agents. In addition, we respond to possible objections and derive several policy suggestions.

1. Introduction

Nudges, subtle tweaks to choice environments that predictably alter people’s behaviour without restricting their option-sets or changing economic incentives (Thaler and Sunstein 2008: 6), are often objected to on moral grounds for their purported autonomy-undermining properties. Hausman and Welch (2010: 128), for instance, object that nudging diminishes ‘the extent to which [individuals] have control over their own evaluations and deliberations’. Nudges, in their view, are no less autonomy-threatening than coercion, especially given their covert form (Hausman and Welch 2010: 130). ¹ Grüne-Yanoff (2012: 636) voices a similar concern about nudges being ‘manipulative … because they deliberately circumvent people’s rational reasoning and deliberating faculties, and instead seek to influence their choices through knowledge of the biases to which they are susceptible’. Rebonato (2014: 371) also criticizes nudges for entailing an ‘effective reduction in decisional autonomy’.

In most treatises on nudging, critics assume the role of nudger to be occupied by the government. Nudging rose to popularity precisely because of the basic insight that the same behavioural influencing techniques used by companies – such as framing, increasing salience and triggering other cognitive heuristics – can be employed by government agents for purposes other than profit maximization. Inspired by this insight, Thaler and Sunstein, the original and most prominent advocates of nudges, have been promoting these techniques as a new form of public policymaking (Thaler and Sunstein 2008). Yet, although the standard autonomy-based objections were raised in response to government nudging, they seem to be morally applicable irrespective of who in fact carries it out.

Thus, assessing the impact of government nudges on autonomy seems insufficient. The policymaker’s infrequent exploration of his new behavioural playground is a drop in the ocean of existing behavioural influences on people. This is why we turn our attention to the permissibility of behavioural techniques from a more common source – the market. In this paper, we argue that the oft-neglected market nudges raise more autonomy-related worries than their public counterparts. The purpose of advancing this argument is not to downplay the threat that government nudges may pose to autonomy, but to point out a potentially greater threat coming from market nudges that can be raised on similar moral grounds. Our main claim is that the institutional structure under which market agents operate provides greater incentives and, in some cases, more resources to nudge individuals, thereby threatening the latter’s autonomy more significantly.

The paper proceeds as follows. First, we analyse why market nudges have been overlooked in debates on the ethics of nudging, and introduce a simple typology that informs our ethical evaluation (section 2). Second, we claim that their use raises significant autonomy concerns (section 3). We show that if critics are worried about autonomy being vitiated by nudges, their worry about market nudges should be greater, since market nudges (1) are not constrained by the principles of government nudging (mildness, sensitivity to people’s interests as they see them, and public justifiability) and (2) are often ‘stacked’ in great numbers that overwhelm people’s autonomy. Our verdict is that the use of nudges in advertising, product selection and market interactions in general is morally far more objectionable than government nudging. Third, we respond to objections that could be raised against our view (section 4). Fourth, we present several policy applications derived from our normative conclusions (section 5). Apart from outright bans, we discuss two sets of regulative strategies in advertising that relate to control of content and of public spaces. The final section concludes (section 6).

2. What are Market Nudges?

2.1. Some conceptual preliminaries

The kind of nudges that critics have in mind when they raise autonomy-based objections to government nudging are so-called ‘Type 1’ nudges (Hansen and Jespersen 2013: 21; Baldwin 2014: 836). ² These trigger or tap into cognitive heuristics, (quasi-)automatic psychological mechanisms that help individuals economize on their limited cognitive resources for careful reflection. To give a few standard examples, a nudger can arrange food in a cafeteria so that people are visually cued into picking healthier options, or she can set up defaults for organ donation, counting on the psychological tendency of individuals to stick to the status quo. Neither of these techniques induce reflection or offer reasons, but aim at steering individuals in more automatic ways.

In this paper, we explore nudges as used by market agents (hereinafter, ‘marketers’), and label them ‘market nudges’. Some may object that we are using the term ‘nudge’ inappropriately here. According to its original conception, a nudge is not only supposed to be ‘easy and cheap to avoid’ (Thaler and Sunstein 2008: 6), but should make individuals better-off by their own lights (Thaler and Sunstein 2008: 10). More often than not, market nudges will not satisfy these conceptual conditions.

This is a terminological concern that is not central to our paper. On our understanding here, nudges are techniques deliberately used to influence people’s behaviour by tapping into or triggering cognitive heuristics. This is what sets them apart from more traditional influencing techniques such as incentives, coercion and rational persuasion. We take nudges to be more or less easily resistible behavioural interventions that can be utilized by corporate and government agents for different purposes (for example, to promote the interests of the nudgee, the nudger or the general public). ³

Those who want to define nudges more narrowly, as easily resistible influencing techniques used by governments to promote the interests of citizens – like Thaler and Sunstein (2008) do in their ‘libertarian paternalist’ framework – can simply imagine that we are using some other term that includes behavioural influences used by other agents and for other purposes. Instead of asserting a ‘correct’ way of using the term ‘nudge’, we stick to it partly for reasons of brevity, and partly to acknowledge the source of the original autonomy-related objections. Our broader consideration of behavioural influences under a more inclusive ‘nudge’ label helps move the discussion away from the ‘paternalism versus autonomy’ debate, considers other ways in which influences can threaten autonomy, and reveals that market influences are in fact more autonomy-threatening than government influences. Limiting our considerations only to nudges, originally conceived, would rob us of this pertinent and pressing ethical analysis.

2.2. Why market nudges have been largely overlooked

The possibility of market nudging is acknowledged in Nudge very early by Thaler and Sunstein, when they argue that choice architects may simply utilize nudges to ‘Maximize profits, period’ (Thaler and Sunstein 2008: 2). The book later sketches how some heuristics can be profited on, without the advancement of the targeted individual’s welfare (e.g. in the credit market; Thaler and Sunstein 2008: 132–144). Thaler and Sunstein go on to claim that:

the invisible hand will, in some circumstances, lead those trying to maximize profits to maximize consumer welfare too. But when consumers are confused about the features of the products they are buying, it can be profit maximizing to exploit their confusion, especially in the short run but possibly in the long run too. (Thaler and Sunstein 2008: 239) ⁴

Still, in Thaler and Sunstein’s defence of government nudging, the purpose of nudging is usually to counteract not the exploitation of biases by profit maximizers, but biases in general. Their emphasis is on the failures of individuals to get things right, and not on the tendency of marketers to exacerbate this.

Similarly, McCrudden and King criticize Nudge and later work by Sunstein for fixating on governments and consumers, while ignoring producers. They state that in ‘prominent examples, Sunstein largely sets to one side the triangular relationship of government, producer and consumer’ (McCrudden and King 2016: 103; emphasis in original). One such example in the literature, that of New York’s former mayor Michael Bloomberg restricting the size of soda cups in the hope of reducing obesity, is explained by Sunstein in terms of people tending to choose drinks in large cups and the government engaging in paternalism to counter their bias (Sunstein 2015 a: 75–76), rather than ‘seeing the soda cup ban as restricting the harms caused by soda manufacturers’ (McCrudden and King 2016: 103; emphasis in original).

Why have nudges from producers largely been overlooked? Multiple factors are likely at play. First, scholars may have simply followed Thaler and Sunstein’s lead and focused primarily on government nudges, since they were such a novel concern. With this shift of academic attention, it is possible that the marketer’s psychological techniques have, at least temporarily, gone off the radar. Second, this focus might stem from slightly ideologized notions about government and business. Sometimes, governments are seen as far more powerful than businesses, and their influence is thought to pose greater threats to freedom and autonomy. Governments are often viewed as huge, monolithic and monopolistic structures that can go unchecked when they nudge, whereas marketers are viewed as much smaller, and their influences scattered and contested by the competition that characterizes decentralized markets. While these views are not shared universally, they explain why laypeople and academics voice their worries when governments discover new tools to exert power over their subjects. These worries are not always felt as strongly about the realm of corporate power.

But this perspective is outdated. Particularly with the rise of new information technologies, large companies such as Alphabet and Meta (owners of Google and Facebook, respectively) tower over many governments of the world in terms of knowledge, capacity, power and resources to steer the decision-making processes of individuals (for a similar point, see Ruehle 2018). And even if we disregard widespread online nudging, it is still the case that much more of our day-to-day choices will be steered by powerful market agents, such as banks, fuel companies or supermarket chains, than by government agencies. Marketers invest significant resources when designing not only advertisement campaigns (targeted or non-targeted), but also products, websites, apps and physical stores, in the hope of influencing the decisions of consumers.

We argue that it is morally erroneous to ignore the activities of marketers in ethical debates on behavioural influences, for four reasons. First, marketers have a much greater cumulative capacity compared with governments to steer behaviour. The vast resources and omnipresence of companies should give us pause regardless of our previous reasons for fixating on government nudges. Second, they are, on average, much more experienced in, and had been familiarized with influencing decision-making long before behavioural findings were picked up by policymakers. Third, marketers have stronger incentives to use nudging techniques. For one, unlike governments, they cannot rely on more coercive measures such as laws or taxes to get individuals to do what they want them to do. All marketers can do is try to inform, persuade, seduce and nudge consumers to buy their products, services, or do whatever makes them money. Fourth, as we will argue in more detail, market nudges are not constrained by principles of benefitting those they target or helping to preserve their autonomy. Yet, they influence people in areas where decisions are considered by most to be non-trivial. As Posner (2013: 212) states: ‘Businesses know, and economists are learning, that consumers are easily manipulated by sellers into making bad choices – choices they would never make if they knew better – in borrowing and investing, and in buying goods and services, such as food, health care, and education’.

2.3. Nudging-to-sell vs. selling-a-nudge

Marketers use behavioural influences in different ways. Our focus in the paper will be entirely on nudges they use to steer consumer behaviour; we will not go into techniques that companies use to influence their employees. Note that the examples below are only a snapshot of a plethora of influences marketers have utilized over the years.

The most widespread kind of market nudging, and centrally relevant for considerations of personal autonomy, is nudging-to-sell. Nudging-to-sell denotes the marketer’s use of heuristic triggers to make it more likely that consumers will pick a particular product or service (for a host of examples, see Ang 2020). The understanding that marketers possess of behavioural influencing may often be non-scientific and cultivated through market experience, but their techniques may often be more effective than those devised by behavioural scientists (even if it is not specified or clear how exactly they exploit cognition or whether they have been tested in controlled environments). ⁵

It is important to distinguish nudging-to-sell from selling-a-nudge, the practice of presenting some behavioural aspect of a product or service as a selling point. Take wearable activity trackers that induce you to exercise regularly by providing salient and timely feedback on your (in)activity. Or take HelloWallet, a finance aid that helps users stick to their commitment to improve savings by appealing to defaults, peer effects and frames. Such behavioural designs are used to convince customers to buy this product or service instead of an alternative without that feature. We do not focus on this category in what follows, since we believe there is nothing wrong with selling-a-nudge. In such cases, a nudge is sold as a commodity to informed and consenting customers.

2.4. Nudging-to-sell in three domains: practices and examples

Before analysing the extent to which nudging-to-sell threatens personal autonomy, we first want to distinguish between three domains of market interactions in which it can be expected. The first is in direct dealings with salespersons. From telemarketing to real estate, salespersons know how to trigger heuristics in verbal exchanges in order to make sales. In car dealership, for instance, salespersons often make some information more or less salient: drawing the customer’s attention on additional but largely irrelevant components, or zooming in on the price of monthly payments, while distracting from the length of payment, fuel prices, etc. (Akerlof and Shiller 2015: 62–63). Or take the practice of haggling, in which salespersons rely on anchoring. The initial price stated by the salesperson will often set the tone for the interaction and influence the final price.

Second, nudging-to-sell also occurs when companies deliberately frame different options, either online, in brochures or in physical choice environments. Take Dan Ariely’s well-known example, which describes three options for purchasing subscriptions for The Economist: web-only for $59, print-only for $125, and a print/web subscription also charged $125. As anticipated, hardly anyone picks the print-only subscription, which is only there to frame the print/web option as ‘a bargain’ (Ariely 2008: 2–3). When the print-only option was included, 84 out of 100 students chose the print/web subscription, while only 16 students picked the web-only option. Without the print-only ‘decoy’, only 32 picked the more expensive print/web deal and the majority (68 of 100 people) picked the cheaper web-only option (Ariely 2008: 5–6). Techniques using anchors and decoys are ubiquitous in market settings. Restaurants, for instance, use decoys in the form of particularly expensive meals, so that slightly less expensive (but still profitable) dishes get picked more often.

Both online and offline, choice environments, i.e. ways in which options and information are framed and presented are inevitable. That said, companies can deliberately design them to maximize their profits. In fact, a handsome amount of resources nowadays is spent on determining exactly how shelf placement best captures the consumer’s eye. ⁶ Or consider the strategies of supermarkets to put temptation items like confectionery, magazines and cigarettes at the counter, while placing milk and eggs in the back so that you have to traverse the entire store in order to get to them (and presumably, buy more items as you go) (Akerlof and Shiller 2015: 21). ⁷

And then there’s the third, and perhaps the most obvious domain in which to expect nudging-to-sell: advertising. Certain forms of advertising have for some time been flagged for their potential to override autonomy (Crisp 1987). Ads effectively utilize heuristics, including positive and negative frames, statistical frames, zero-risk bias, salience, priming, the bandwagon effect and anchoring. Akerlof and Shiller (2015: 45) note that advertisers have not only been honing their skills at utilizing behavioural techniques, but have developed sophisticated methods to measure their success. ⁸

To further stress the importance of this analysis, we should realize that these techniques are not aberrations to how markets work. In fact, one may wonder whether markets are conceivable without them. Market interactions seem impossible without salespersons, advertising and deliberately presenting (information about) different choice options. In all of these cases, companies should be expected to be biased towards selling. If we are not arguing against markets themselves, which are more reliable than other economic systems at producing efficiency and economic growth, what is the point of criticizing market nudges that markets simply cannot do without? ⁹ Furthermore, even if it is conceded that advertising indeed has the capacity to undermine autonomy (which we will argue below), it might only be pro tanto wrong given that the harm can be outweighed by its positive social function to inform consumers and thus increase their choice range.

These are all noteworthy points. The success of any normative argument we make here about the capacity of the market to undermine autonomy would have somewhat indeterminate policy implications. Our policy decisions may often favour current market practices, if alternatives only seem worse or more uncertain. For example, if we cannot yet think of ways to regulate the interaction between salespersons and consumers without paralysing business operations, then we should hold off on regulation. Still, we can respond to these worries by making two points. First, some policy ideas are available for tackling at least some heuristic exploitation. We might think that the way supermarkets or option menus are arranged should not be completely at the discretion of the seller. We might also think that the ways in which we organize public spaces determines, at the very least, the extent to which the population is exposed to advertising; ousting advertising from some public spaces, for instance, would seem an available policy option. Second, it would be normatively sufficient to offer a strong autonomy-based argument against market nudges, and claim that it should guide future policy engineering. The argument would not be undermined just because there is no readily available policy formula. These two points do not have the capacity to protect consumers from autonomy-undermining influences entirely, but could shorten the range of threats, which consumers can then overcome more easily. We will tackle these concerns more extensively in section 4 of the paper and will now turn to detailing our argument pertaining to autonomy.

3. How Market Nudges Threaten Personal Autonomy

We turn now to outlining the threat that market nudges pose to personal autonomy. We test whether market nudges disrupt or undermine the processes of, on the one hand, setting our ends and value commitments, and on the other hand, pursuing them (Engelen and Nys 2020). Following DesRoches’ arguments, we start from the assumption that individuals (1) have at least some value commitments that are ‘normative, affective, and stable’ (DesRoches 2020: 563) (for instance, a person may be committed to her health) and (2) will derive preferences from these value commitments. Market nudges, we will argue, have the capacity to reduce the control of affected agents over the formation of these preferences.

3.1. Considerations of personal autonomy

While there are many conflicting accounts of personal autonomy, quite a few of them revolve crucially around the control agents have over the formation of their preferences. According to these accounts, autonomy concerns how agents make such choices and whether they, in some sense, are in control. To specify this vague notion of control, we take cue from Christman’s two conditions for personal autonomy:

1. (i) Person A acting on a preference P is autonomous if A did not resist (or would not have resisted) P’s development when attending to this process of development;

2. (ii) The lack of resistance to the development of P did not take place (or would not have) under the influence of factors that inhibit self-reflection. (Christman 1991 a: 11)

Take someone who enters a cafeteria and buys a piece of confectionery. How did her apparent preference for something sweet come about? Did the nudge-enhanced design of the cafeteria interfere with the formation of the preference? According to Christman, the person is autonomous if nothing inhibits her capacity to reflect on her preference formation, and if she would not have resisted the formation of her preference had she attended to it.

According to the objections to government nudges posed at the outset, nudges may arguably inhibit self-reflection, and undermine the capacity of individuals to resist the development of a preference as a result. Notice, however, that Christman’s account allows that people willingly subject themselves to nudges while retaining their autonomy. First, they may attend to but not resist how these nudges shape their preferences. Second, if they do not attend to nudges influencing their preference formation (which happens a lot), their autonomy is only undermined if they, hypothetically, would have attended to and resisted this formation process.

Not resisting the development of a preference and manifesting only minimal rationality in the process of self-reflection is fully compatible with Christman’s view on autonomy (Christman 1991 b: 347). Christman’s autonomous agents are not expected to assume full reflective control, making his account plausible in light of behavioural findings. Christman (1991 a: 20) acknowledges ‘that people often subject themselves, in ways and under conditions that manifest autonomy, to factors and influences that severely undercut their reflective capacities’. In short, these two key capacities – being able to attend to and resist the development of a preference – preserve autonomy, but autonomy is not necessarily undermined if agents decide against or neglect to pay attention to their preference formation processes. ¹⁰

People may not pay this kind of attention, for instance, if they do not care about being nudged in the market setting, if they believe their particularly strong preferences will remain unaffected, ¹¹ or if they would rather save up their limited ‘mental bandwidth’ (see Mullainathan and Shafir 2013) for reflecting on decisions they deem more important. For many, the market is exactly the setting where they wish to invest little cognitive effort. ¹² Customers in a supermarket or restaurant are often willing to go along with the nudged choice environment because many decisions there have small stakes for them.

While we do believe that our personal autonomy is generally safeguarded when we allow ourselves to be steered in market settings, we should be careful not to assume too readily what decisions individuals deem important, and to what extent they would accept influences by third parties interfering with their preference formation. ¹³ In order to assess whether and when market nudges lead to preference formation processes that threaten the autonomy of consumers, we believe we need to focus more on the latter’s underlying ends and value commitments, which are more fundamental than the preferences they form on the fly.

We attend to the above two autonomy-related processes in reverse order. We first evaluate the threat of market nudges to autonomously pursuing our ends and value commitments, and then turn to the more mysterious relation between market nudges and autonomously setting those ends and commitments in the first place.

3.2. First threat: market nudges can make people deviate from their ends

Objections to government nudges that we mentioned at the start of the paper state that nudges can bypass reasoning, and could thereby diminish the control individuals have over their deliberations, evaluations and actions. Given that market nudges seem to trigger the same kinds of cognitive processes as government nudges (bypassing reflection and reducing control), market nudges seem just as worrying as government nudges with regard to autonomy. We call this the similarity argument.

We claim, however, that market nudges turn out to be much more worrying for personal autonomy than government nudges. This a fortiori argument, as we call it, rests on two counts, which we elaborate in the following two sub-sections.

3.2.1. First count: principled constraints on government nudging do not apply to market nudging

First, in the absence of regulation, market nudging is not constrained by the same principles that regulate government nudging in liberal democracies like ours. Whenever government or public agencies implement nudges, in ways envisioned by Thaler and Sunstein (2008: 6), these are standardly constrained by the following three principles: (1) resistance-enabling mildness (cheap and easy to avoid), (2) a sensitivity to nudgees’ views about their own interests (promoting what nudgees themselves judge is best for them), and (3) public justifiability. While these principles constrain government agents engaged in nudging, we stress that, importantly, they do not inform market agents in devising nudge strategies. Let us go into each of these constraints.

As for the first, governments employ nudges as one of many tools at their disposal (informing, campaigning, incentivizing, sanctioning, coercing, etc.). If they believe nudges lack effectiveness in influencing people’s behaviour, they may often justifiably revert to legal stipulations, sanctions, taxes, etc. But when they do nudge, most proponents agree governments should make sure that nudges are easy and cheap to avoid. To achieve this, some believe that governments should employ nudges transparently, so as to make citizens watchful (Bovens 2009; Ivanković and Engelen 2019) or to reduce the control governments have over citizens (Schmidt 2017).

By contrast, since marketers cannot coerce consumers into buying their products or services, they have an incentive to steer consumers in ways that go unnoticed but have a bigger impact than people might expect and want (were they to attend to the formation of their preferences). While advertising is often presented as a means of informing consumers (Bagwell 2007), it is typically enhanced by whatever psychological tricks marketers have up their sleeves. Moreover, marketers have strong reasons not to make consumers alert to nudging at all. In contrast to policymakers, they will not give people hints about how they are using their techniques, ¹⁴ or subscribe to any rule that prohibits exploiting their cognitive weaknesses. In short, given the circumstances they operate in, marketers will typically have no reason to ensure that their nudges are mild and easy to resist.

In addition, government and market nudgers also differ with respect to their commitment to promote people’s interests, as people themselves understand these (second constraint; interest sensitivity), or to promote the public interest (third constraint; public justifiability). As a matter of principle, governments that nudge should at least try to serve the interests of their citizens, either directly by paternalistically making ‘choosers better off, as judged by themselves’ (Thaler and Sunstein 2008: 5), or indirectly by serving the public interest (Guala and Mittone 2015; Moles 2015).

While this may paint an optimistic picture of governments, it is at least how government nudges were envisioned by Thaler and Sunstein ¹⁵ and how they have been justified on a democratic basis (Sunstein 2015 b; Schmidt 2017). In addition, institutional mechanisms are in place that hold liberal democratic governments accountable and incentivize them to act on these principles. Think here of fair, free and regular elections, legal and deontological codes for public servants, the public and transparent nature of discussions and decisions made by political bodies (which incentivize politicians not only to talk the talk – and defend positions that they deem publicly justifiable – but also to walk the walk ¹⁶ ), public scrutiny from independent organizations and a free press, etc. When governments design so-called ‘dark nudges’ (Petticrew et al. 2020) that are at odds with people’s interests (in their own lights) and/or lack transparency or public justification, they will likely be called out and possibly face consequences, such as electoral failure. While nudges implemented by governments can and sometimes do undermine people’s autonomy by bypassing reflection and reducing control, policymakers and bureaucrats have good reasons to ensure that people can notice and resist such influences (first constraint; see also Ivanković and Engelen 2019) and that nudges are in line with people’s interests (second constraint) or publicly justifiable in some other way (third constraint).

To be clear, we do not deny that policymakers and bureacrats can (be incentivized to) employ nudges that ‘realize their own intermediary goals (on their way to re-election)’ (Schubert 2017 a: 515). They may nudge to serve their own interests instead of the public good and may do so covertly instead of transparently. Starting with Buchanan (1987), public choice theorists have stressed that political behaviour is as self-interested as market behaviour and that there is no reason to assume that government agents are any more benevolent or public-spirited than market agents (Tresch 2015: 15–17).

Obviously, governments have their own interests that they can try to promote by nudging individuals in different ways. Electoral cycles and other aspects of the institutional framework under which they are operating (like a commitment to economic austerity) can incentivize them to opt for nudges instead of less popular and more costly measures such as subsidies or the enforcement of legal constraints. In the nudge literature, this point has forcefully been made by Schubert (2017 a) as part of his ‘political economy of nudging’. In fact, it dates back to Jolls et al. (1998: 1545) who already claimed that ‘behavioural bureaucrats’ can cause government interventions to ‘make things worse’ rather than better.

Three remarks are in order here. First, the extent to which government agents actually employ hard-to-resist, self-serving and publicly unjustifiable nudge policies is an empirical question that cannot be decided on the basis of blanket generalizations or assumptions (which can be too naive but also too harsh). Policies should be informed by empirical findings about the biases, preferences and motivations of those being nudged and of those doing the nudging (in both markets and politics). Second, as mentioned, our main claim and argument here is comparative in nature, namely that government agents are governed by principles and constraints that are largely absent for market agents. The three aforementioned constraints can be expected to govern public policymaking in well-functioning liberal democracies in ways not applicable to market agents. When public agents use nudges that are hard to resist, insensitive to people’s interests and/or lack public justification, they will likely be blamed for violating principles they are supposed to uphold. Companies doing the same may face backlash as well, but are, importantly, not expected to uphold such principles in the first place. In other words, if government nudges raise concerns about undermining personal autonomy, market nudges should raise greater concerns in this respect. Third, our approach is in fact in line with Buchanan’s: instead of assuming that government agents are more benevolent than market agents, we claim the main difference lies in the institutional framework in which they operate. ¹⁷

Because the above three principles do not regulate market settings, market nudges are more likely to be detrimental to people’s autonomy, i.e. to the level of control people have over their preference formation processes. Marketers are simply not constrained in the same ways that policymakers are. Surely, they may take mildness, interest-sensitivity and public justifiability into account, but only when and to the extent that this coincides with their economic incentives. And while there are some legal regulations in place, such as forbidding fraud, there are no principled reasons and no corresponding institutional mechanisms for market agents to constrain themselves to easily resistible nudges that have people’s private and public interests in mind.

Market nudges do seem more tolerable when they prod people to act in line with their value commitments. Many market nudges are designed with such purposes in mind. Following Beggs (2016: 127), market nudges that both maximize profits for marketers and promote the interests of consumers can be called ‘Pareto nudges’. Conversely, some techniques exploit biases in order to maximize profits for marketers, but run counter to the benefit of consumers (Beggs 2016: 138). Call these exploitative nudges. ¹⁸ Although Pareto nudges are not constrained by mildness in principle, and may influence individuals when they would rather not be influenced, they are at least more respectful of personal autonomy than exploitative nudges. If the market provided enough incentives for marketers to carefully gauge the value-based preferences of customers and stick to Pareto nudging, market nudges would raise fewer autonomy worries. While companies often do aim to engage in Pareto nudging, they will only do so when this fits into a broader profit strategy. Sadly, it will often be more profitable for marketers to tempt weak wills instead, and lure people into decisions they have reason to avoid and regret. In so doing, they undermine the synchrony between individuals’ first-order preferences, as shaped by the exploitative market nudges, and their underlying value commitments.

Some will no doubt suggest that a healthy competitive market will punish such exploitation. If consumers’ cognitive weaknesses are exploited by marketers, they would claim, consumer demand will go to other, more trustworthy suppliers. Sugden (2018), for example, argues that the market is essentially about mutual advantage, offering participants both opportunities and incentives to mutually benefit each other. Businesses will not make profits if they fail to provide products and services that consumers are willing to purchase (and that willingness arises from the perceived advantage). Even if consumers are initially misled, claims Arrington (1982: 8), the product is unlikely to be purchased again, forcing businesses to withdraw deceptive advertisements. This boils down to claiming that competitive markets incentivize Pareto nudges and restrict opportunities for exploitative nudges.

However, the expectation that businesses will stick to using Pareto nudges and that consumers will flock to these businesses is too optimistic about consumer behaviour. Consumers may sometimes forget about a particular purchase. They may also fail to notice that they have been nudged by marketers, and even if they do notice, they may not recognize this influence as decisive in producing their actions. Their failures in both cases may be caused by an undue optimism that they are immune to market influences such as advertising, or at least significantly more resistant than others (Berger 2017). In addition, consumers have been shown to rationalize post hoc whatever they ultimately decide to do, which is known as ‘post-purchase rationalization’ (Mather et al. 2000). This casts doubt on whether consumers would end up blaming nudgers for bad decisions, and not themselves.

Generally, nudges can be said to run across three spectrums: (1) easily versus hardly resistible, (2) more or less sensitive to people’s interests, and (3) more or less publicly justifiable. Our claim is not that government nudges occupy the positive side of each spectrum, whereas market nudges occupy the negative side, and our aim here is not to redeem government nudges. Rather, we suggest that the placement of nudges on these spectrums mostly depends upon institutional mechanisms that are in place when liberal-democratic governments consider nudging, but that are largely lacking when it comes to companies. While governments may try to implement forceful and self-serving nudges that are not publicly justifiable, these mechanisms will help detect and criticize those practices, as they are clearly at odds with how liberal democratic governments justify themselves. Companies, on the other hand, are not faced with the same kinds (or the same level) of demands for justification.

3.2.2. Second count: nudge stacking

The second, and possibly graver count of the a fortiori argument, on which market nudges seem far more threatening to autonomy than government nudges, is what we call ‘nudge stacking’. Coons and Weber (2013: 21–22) have pointed out that, in any given choice environment, there could be more or fewer nudges at play, and if there were more, this would likely constitute a stronger influence on individuals. To the extent that government choice architects in liberal democracies uphold the aforementioned principles, they have reasons to refrain from excessive nudging. After all, the worry here lies not with this or that nudge or with the underlying end it promotes, but with their sheer volume. While the exact effect of increasing the volume of nudges is uncertain (e.g. perhaps some nudges would cancel each other out) and should prove a fruitful area for empirical investigation, it seems at least likely that the more nudges there are (in any particular choice environment and across different environments), the more personal autonomy is under pressure (ceteris paribus). The more environments are flooded with nudges, the less likely it seems that people will successfully navigate and resist all of them, even if every single nudge is mild.

In market settings, nudges seem likely to be excessive in number. Competitive markets produce strong incentives for the many agents operating in them to nudge if they want to keep up with competitors. In fact, it seems at least that the healthier the market competition, the more marketers there will be to engage in nudging, making it more likely that market nudges will be stacked and, consequently, the more difficult it may become for consumers to navigate the behavioural landscape.

In addition, interactions between marketers and consumers are increasingly occurring online. The advancement of new and expanding communicative channels provides new opportunities for effective market nudging (Yeung 2017). If marketers indeed enjoy a continuously expanding platform for behavioural influencing, and are incentivized in doing so to keep up with their competitors, then market nudges are likely to become stacked, which poses the threat of additive harm. In other words, even if a single market nudge (one subway billboard, one deliberately designed search result, one physical store) does not gravely threaten autonomy, the more significant threat may come with their ever-greater numbers. To employ so much nudging also gives marketers ample opportunity to polish their methods. Abdukadirov (2016: 182) has offered an effectiveness-based claim that markets are better for coming up with and updating nudge techniques than government policy, given the greater feedback from targeted individuals. As a result, however, autonomy might be impacted to a greater extent. If stacking nudges will upgrade the marketers’ know-how for steering behaviour, then this will also increase their capacity for undermining autonomy.

Finally, market nudge stacking raises concerns not only because it threatens consumer autonomy, but because different people’s autonomies may be affected very unequally. Mullainathan and Shafir (2013) have argued convincingly that scarcities of all kinds occupy people’s attention and reduce their capacities for self-control. Because the attention of the poor and the deprived is occupied by more than one of such scarcities, they generally find it harder to resist temptation. Since market nudges are often designed to tempt weak wills (or in the case of the poor – weakened wills), their effects may disproportionately threaten the autonomies of the poor. If this turns out to be empirically correct, then allowing uninhibited market nudging would raise justice concerns on top of autonomy concerns.

We should reiterate that the empirics are still lacking. Neuroscientists and behavioural scientists should investigate how our brains process environments stacked with market nudges. Do we suppress some of the advertising into the background and succeed in not noticing it? Or do we become overwhelmed, so that greater impact results from greater numbers? These questions caution us against making hasty judgements about how exactly nudge stacking relates to personal autonomy.

That said, we want to make three careful, yet commonsensical suggestions. First, nudge stacking in markets seems relevant for autonomy because it increases the likelihood that individuals will be steered in directions that go against their values and ends. After all, the more nudges there are in the market, the more likely it seems that some effects will go unchecked. Second, an abundance of market nudges entails more temptation in areas where individuals often feel weak (think of our predilection for sugary and fatty foods and drinks that fast-food marketers exploit). Crucially, our attempts at resisting these nudges, through learning where to anticipate them and being on the lookout for them, will require cognitive effort (or mental bandwidth) that individuals can no longer expend on other purposes (Mullainathan and Shafir 2013). Third, nudge stacking is likely to have effect not only on whether we succeed in pursuing our ends and value commitments but on how we set them. We turn to this in the following sub-section.

3.3. Second threat: market nudges can unduly intervene in people setting personal ends

So far we have established that market nudges can inhibit consumers in pursuing their values and ends, for example by playing into their weakened wills. Here, we ask whether there is a more fundamental kind of influence that marketers can have, namely whether they can interfere in how people come to adopt their underlying value commitments and ends (Engelen and Nys 2020). If market nudges succeed in changing our behaviour, they may be able to do the same to our values and conceptions of the good life (see for example Schor 1998; George 2001; Lindstrom 2011). If marketers successfully nudge us into trying out a new fast-food restaurant (or any brand of car, clothing, etc.), we may end up, well, loving it.

Discovering and acquiring new tastes can be perfectly compatible with consumer autonomy. However, consumers may also be unaware of and unwilling to acknowledge the extent to which market nudges non-reflectively influence their conceptions of the good life. Thus, it is possible that they come to adopt some value or end but would, hypothetically, have resisted this formation process had they attented to it.

While the influence of market nudges on day-to-day preferences, when entering supermarkets or shopping online, is fairly straightforward, the process of setting one’s values and ends is much more intricate. As a result, justifying the claim that market nudges unduly intervene in this process requires more than just showing that market nudges take part in changing people’s values: one should also show that they have made it very difficult for consumers to reflect upon the formation of their values and ends.

As a general claim about nudges, this is too strong. While we are surely nudged to buy stuff we do not need (Schor 1998), typically this does not inhibit us to take a break every now and then, ask what it is that we need and set our own priorities in life (Engelen and Nys 2020). In specific instances, however, this claim is very plausible. First, think of nudges that trigger or play into addictions. As Schüll (2012) details in her ethnography Addiction by Design, casinos are deliberately designed – using physical architecture, colors, and sounds – to ensure that people keep gambling. Such ‘dark nudges’ (Petticrew et al. 2020) play into people’s addictions and vulnerabilities without enabling control or reflection over them. Second, market nudges, especially when stacked, can induce a lot of small changes in people’s behaviour, that in the long run and in combination with other non-reflective processes such as cognitive dissonance reduction, lead to non-autonomous changes in people’s values (Engelen and Nys 2020: 149). While it is hard to imagine how a simple market nudge could manipulate us into revising our fundamental value commitments, a more subtle, indirect and long-term impact of a flood of nudges steering us in one direction is surely conceivable.

4. Objections

We have claimed that market nudges have the capacity to undermine personal autonomy, and may in fact pose a more considerable threat to it than government nudges. This claim is not without opposition. First, we address the claims of White (2013), whose arguments run in the opposite direction to ours and who claims that market nudges, unlike government nudges, are not disrespectful (4.1) and are easier to expect (4.2). We then attend to further objections to our claim that market nudges should be constrained, namely that this is too demanding (4.3), inconceivable (4.4) and that market nudges do not gravely infringe personal autonomy (4.5). In our view, these objections partly explain why nudge theorists have shied away from ethically evaluating market nudges. Addressing these objections will also inform our short overview of policy suggestions in section 5.

4.1. Market nudges are not disrespectful

In his book The Manipulation of Choice, White (2013) offers a moral claim opposite to ours – that government nudges are by far more objectionable than market nudges. White’s reasons for favouring market influences build on his prior objections to government influences. At the centre of his objections is the so-called knowledge problem, referring to the inability of government agencies to track the interests and values of citizens (as they themselves understand them), and, thus, to nudge them towards options these citizens would choose by their own lights (see also Rizzo and Whitman 2009). Since governments lack knowledge about people’s interests and values, any government nudging aiming to promote these is presumptuous about their content. This is why, White believes, government nudging is marked by a disrespectful attitude that offends against the personal autonomies of its targets (White 2013: 117).

As we noted earlier, however, there are both normative principles that governments are expected to uphold and institutional mechanisms that incentivize them to do exactly this – nudge only in easily resistible, interest-promoting, transparent and publicly justifiable ways. When they do, government nudging would not be so obviously offensive to autonomy, since people’s capacity to consent to or reject the presumptions of government would be strengthened. But let’s imagine, for the sake of White’s argument, that such transparency is not feasible and that governments will do at least some tenuous approximations of people’s interests and values. White believes this is offensive, or at least more offensive than how businesses operate:

Although the government … presumes to promote your interests, businesses have their own interests – primarily, to maximize profit. Businesses are interested in your interests only insofar as those interests lead you to buy their products, not for your good. Whatever you think of the profit motive, the advantage of the single-minded purpose behind much business behavior is that they’re not presuming to make decisions for their customers in their own interests. (White 2013: 108–109)

Assume that White’s claim is always true – governments always nudge in line with their approximation of the nudged person’s interests and values, while marketers nudge without regard for those and for their own benefit. What makes the first attitude disrespectful, as compared with the second attitude? White is not explicit about this. He explains that the first attitude is disrespectful because it is presumptuous about the nudgee’s interests and values. But would it follow that the second attitude is respectful, because it is, presumably, not presumptuous? Attitudes can be disrespectful for other reasons.

While one could argue that behavioural influencing is known, by consumers, to be part of the marketers’ modus operandi, and that marketers are honest about their profit-seeking motivations, their attitude might still be quite disrespectful in its own way. It is characterized by an objectionable indifference to the value of citizens’ autonomous pursuits compared with one’s own, and the readiness to use heuristic triggers to gain at their expense without providing them with any help to opt out. Arguably, this seems a disrespectful attitude. Once again, marketers will often have an incentive to nudge customers in line with their perceived interests and action plans, but this is entirely contingent upon their short-term or long-term profit strategy. The regard for the customer’s interests and values is purely instrumental for marketers within a market rationale.

Additionally, even if motivational honesty about one’s ends somehow were sufficient for being respectful towards nudgees, it is not obvious that marketers will predominantly be the honest, whereas governments the dishonest party. We agree with White that, as a rule, companies are in the business of making money. But this also means they are willing to go the extra mile to distract their customers from this general fact – by convincing them that they ‘care about’ and ‘respect’ whatever they think their customers care about (their own welfare, environmental benefits, or what have you). As for government agents, whether we expect them to be honest about their motivations will often depend on the level of trust we put in governments. In liberal democracies, such expectations of government agents will tend to be higher, but so will the institutional pressures placed upon them to uphold transparency about their motivations and their policies. In these circumstances, it is reasonable to expect governments to be pressured into motivational honesty. Hence, White’s notion of honest businesses and dishonest governments is exaggerated.

4.2. Market nudges are easier to expect

Another suggestion by White for why market nudges might be less detrimental to personal autonomy is that, unlike government nudges, we learn to expect them:

We know businesses do whatever they can – hopefully within the law and other ethical norms of their industry – to make money, and we fully expect them to do whatever they can to make money, even if we’re not aware of what exactly they do. It doesn’t take an extreme belief in caveat emptor (‘let the buyer beware’) to realize that consumers should expect businesses to manipulate their behavior to some degree and to guard themselves against it to whatever extent they can. (White 2012: 109–110; emphasis in original)

Expectations of consumers seem to be relevant for the preservation of autonomy in the face of interferences that could divert them from pursuing their ends. If we expect to be nudged in the market setting, this can help us fend off some autonomy-undermining influences. But this general expectation is not sufficient for protecting autonomy in most instances. Transparency about the fact that nudging techniques can and will be used – what Bovens (2009: 216) has called type interference transparency – is insufficient, for it tells us very little about the specific kinds of techniques in use, in what way these techniques are deceptive, how numerous they will be, whether they are resistible, etc. For instance, as Bovens (2009: 216) states, governments could be transparent in this general way about the use of subliminal advertising, but that would hardly render it permissible.

In order to protect their autonomies, consumers need to be able to access the more fine-grained information about market nudging techniques. As we have argued before, marketers will be extremely reluctant in providing this information (in contrast to democratic governments where such a transparency constraint is more likely to be respected).

4.3. Refraining from market nudging is too demanding

Is it overly demanding to require marketers to refrain from behaviourally influencing prospective consumers for profit-maximizing purposes? Thaler and Sunstein (2008: 5) ‘argue for self-conscious efforts, by institutions in the private sector and also by government, to steer people’s choices in directions that will improve their lives’. However, according to Sunstein, circumstances overriding this plea are quite common in the market setting:

In identifiable cases, those who do not exploit human errors will be seriously punished by market forces, simply because their competitors are profiting from doing so. Credit markets provide many sad examples. Consider cell phone plans, credit card plans, checking accounts, and mortgages … In all of these areas, it is possible that companies that provide clear, simple products would do poorly in the marketplace, because they are not taking advantage of people’s propensity to blunder. (Sunstein 2015 a: 10–11; emphasis in original)

Thus, in certain areas of the market at least, requiring marketers to refrain from for-profit influencing would be tantamount to asking them to allow their businesses to go under. While there might be areas of the market in which there are incentives for such restraint (for instance, to preserve the company’s reputation), Akerlof and Shiller (2015: xii) state that:

unregulated free markets rarely reward … heroism, of those who restrain themselves from taking advantage of customers’ psychological and informational weaknesses. Because of competitive pressures, managers who restrain themselves in this way tend to be replaced by others with fewer moral qualms.

We concur with both Sunstein and Akerlof and Shiller that, in most circumstances, markets are unforgiving arenas, where staying in the race often demands employing the entire toolset of legally available means. In such circumstances, a requirement of restraint seems overly demanding. If, in the near future, people become acutely sensitive to the exploitation of bias for market gain, say, because it brings about costly negative externalities, marketers may become more likely to respect pleas of restraint (although this will more likely be motivated by business incentives).

A different matter is whether it would be demanding, for both institutions and citizens, to come up with a regulatory scheme that would curtail the use of market nudges. Of course, some legal rules that regulate the use of specific advertising techniques (e.g. forbidding deception) and other design aspects of work environments (e.g. providing safe and healthy work environments) are already in place. That said, a stricter legal framework constraining market nudge use with an eye to protecting consumer autonomy would constitute a new game for companies to play. Note that such schemes need not require the market to be regulated across the board.

For example, restricting the use of market nudges in particular public spaces could be one form of regulation. It would not be counted against it that not all nudges in the market setting are restricted. In fact, a limit to government interference may well be justified. Namely, governments would be stifling marketers if they were interfering with every segment of their business operations. But some regulation could make an important difference given the stacks of nudges that overwhelm autonomous decision-making. Hence, a regulation strategy to protect personal autonomy by restricing some market nudging need not be overly encompassing, nor overly demanding. It can focus, for example, on the most autonomy-threatening aspects of market interactions in areas of decision-making that are most commonly regarded as non-trivial.

4.4. Markets without market nudging are inconceivable

We mentioned in section 2.4 that it might be hard to even conceive a market properly ‘purged’ of market nudging. Practices of advertising and other market nudging have become so hardwired into our practical conception of the market that it is arguably difficult to envisage markets without market nudging, or to assess whether they would be desirable. ¹⁹

The conceivability concern has two versions. First, it might be suggested that market nudges are a constitutive part of market activities. In effect, any argument against market nudges would be an argument against markets themselves. If so, opponents could dismiss our arguments because, although market nudges might undermine personal autonomy, we cannot do without them if we want an efficient system of economic coordination. However, we are not interested in mounting an all-out attack on markets on the shoulders of a criticism of market nudges or engaging in a broader behavioural welfare economics debate about the justification of markets in light of behavioural findings (see, for example, Sugden 2018). Imagine that some market nudges were shown to be constitutive parts of the practices that contribute to mutual advantage to an extent much greater than any other system of economic coordination could secure. This would be sufficient for preserving this space for market nudges. But this will not hold across the board. Some forms of advertising or bartering practices used to swindle buyers are not constitutive, and markets are perfectly conceivable without them. ²⁰ As long as we are able to identify such practices and devise regulatory counters that do not undermine the market as a whole, criticisms of market nudges can be mounted independently of criticisms of markets.

Second, it might be suggested that market nudges are a constitutive part of valuable market practices. For instance, some may view certain aspects of barter in a romantic light, and would object if these practices were eliminated through regulation. They might try to draw on arguments like that of Cohen (2012: 149), who claims that we have good reasons to conserve things that are valuable, even if things of greater value could replace them. If certain market practices involving market nudges indeed have value in terms of culture or tradition, then we might consider placing some constraints on autonomy.

However, a conservative argument truly inspired by Cohen’s account would have to show that preserving practices involving market nudges does not also preserve injustice (Cohen 2012: 144). Rampant market nudging has the capacity to produce profound inequalities. Because these considerations divert us into the domain of justice, we only drop two small hints as to how these troubling inequalities come about. First, as we argued before, rampant market nudging is unjust because it affects the autonomies of the poor more significantly, in terms of increasing the cognitive burdens imposed on their mental bandwidth (Mullainathan and Shafir 2013). Second, it might give marketers a systematic advantage over consumers. This is because marketers in an arena of unfettered nudging are provided with numerous opportunities to test influences and maximize their effect. As a result, consumers, even when they are educated about the workings of nudge techniques, are faced with ever greater challenges to keep track of all the ways in which their heuristics are being targeted.

4.5. Pervasive influences do not undermine autonomy, so why would market nudges?

Contrary to our suggestion about the additive harm of stacked market nudges to autonomy, some might think that the invasion of market nudges in fact downplays autonomy-related objections against them. If there are so many market nudges around, then what difference do a few additional ones make? Better yet, if behavioural influences triggering the same heuristics are pervasive in all areas of social life (Shafir 2016), and if these seem fairly innocuous, then market nudges might not be deeply or gravely autonomy-undermining. Consider Sunstein’s claim:

Much of modern advertising is directed at System 1, with attractive people, bold colors, and distinctive aesthetics … Doctors, friends, and family members (including spouses) often do something quite similar. Is romance an exercise in manipulation? Maybe so. Is medical care? Is the use of social media? A great deal of conduct, however familiar, can be counted as manipulative in some relevant sense; but it would be extreme to condemn it for that reason. (Sunstein 2016: 60)

Sunstein’s point is that if we were to highlight manipulative tendencies across the vast areas of human interaction, the wrongdoing of many single instances of influencing would seem downplayed and would not warrant expressions of condemnation. But his conclusion seems hasty. The pervasiveness of certain practices of manipulation, even those in our private life, and our being used to them, should not be sufficient to think that condemning these practices is extreme.

We do not have the space here to offer a moral assessment of ‘nudging in the private sphere’. Yet, our position can be gathered from our account on market nudges. If agents in either the private or the market setting deliberately aim to trigger our heuristics in ways that push against our value commitments, then we are justified in raising autonomy complaints. The pervasiveness of these practices does not mean that they are unobjectionable. ²¹

Yet, a related concern might arise here. Although ‘nudges’ in our private lives are pervasive and could undermine our autonomy in a similar way to market nudges, we would probably consider state interference inappropriate in our private domain. If so, is state interference appropriate for the market setting?

We believe it is. According to Rawls (1996: 266), markets are part of the ‘basic structure’, the purpose of which is ‘to secure just background conditions against which the actions of individuals and associations take place’. The structure of the market is supposed to fit into these just background conditions. It could be plausibly stated, as we have hinted in section 4.4, that the exploitation of cognitive heuristics in market interactions erodes the background conditions of justice. For instance, the systematic exploitation of heuristics would favour marketers over buyers and allow autonomy to be undermined for market gain. If this would indeed be the case, we would have at least pro tanto reasons to derive legal rules from the basic structure that would remedy the resulting deviations from justice. ²²

5. Policy implications

Now that we have established that regulating the market’s behavioural landscape is appropriate, we turn to considering the policies that could alleviate the threat that market nudges pose to autonomy. Recall that our considerations are aimed at curtailing the use of market nudges to protect autonomy, not rejecting the market or revolutionizing it to the point of seriously undermining the incentives that drive its efficiency. We should be careful not to throw the baby out with the bathwater. Many market nudges are deeply integrated into market interactions (such as product placement) and it would be difficult to conceive the market without them. Thus, a policy scheme should not start an inquisition against market nudges, but should rather look for ways of thinning out their numbers and effects when this is conceivable and appropriate all things considered. The goal of policy should be to gradually reduce the burden on autonomy imposed by market nudges.

Oliver’s ‘budge’ strategy, which focuses on providing public officials with ‘an education in behavioural economic concepts’ and ‘insights into where and how their citizens’ cognitive limitations are being exploited excessively’ (Oliver 2013: 698), may provide the first step towards a successful policy scheme. The goal of the budge strategy is to hamper the ‘profit-oriented industry’ by informing regulators’ decisions ‘on where and how to regulate’ (Oliver 2013: 698–699). So, how should ‘budging’ translate into policy?

One suggestion might be to ‘meet fire with fire’ (Sunstein 2016: 63), by counteracting market nudges with government nudges. After a decade of interdisciplinary research into nudges, government agencies have acquired significant capacities to steer individuals towards desirable options. Why not steer them away from undesirable ones? An informed policymaker could consider where and when market nudges can undermine autonomous choice the most (e.g. choices pertaining to health, finance and big purchases), and devise government nudges that help consumers avoid detrimental decisions.

However, the addition of government nudges would exacerbate instead of address worries about nudge stacking, since choice environments would become even more saturated with nudges. Another difficulty with this strategy is that marketers are better placed than the government to quickly and effectively employ new nudges, and thus counteract the counteracting by government. In short, a nudge bidding war would hardly relieve autonomy concerns.

A policy scheme thus requires counteracting not only particular market nudges, but also nudge stacking and the marketer’s capacity to respond to government interventions. We consider two sets of strategies here, one related to control of content, and the other to control of public spaces. Note that all the strategies we mention here concern advertising, but similar policies could be devised with respect to other nudging aspects of market interactions, such as the regulation of market spaces or pricing.

From the set concerning the control of content, we mention three strategies. First, consider that some ads nowadays hold no actual content about the product or service that they are promoting, unlike the traditional (and justified) conception of ads as informing potential customers. It would not be too controversial to argue that regulation on providing some minimal content should be required, in order to reduce the distraction of the consumer from the relevant information. Second, we noted in section 3.2.1 that marketers will not be too keen to reveal their techniques, but regulation might conceivably require that their methods be disclosed. This would allow consumers to look up market techniques in areas of decision-making that they deem relevant, in order to more successfully pursue their values and ends. Third, one common technique in recent years has been to disguise commercials as news. Since consumers often fail to spot the difference between news and commercials (e.g. in the form of ‘informercials’), and thus fail to be on the lookout for market nudges, regulation could require that commercials be clearly marked and distinguished from news pieces.

The second set of strategies concerns the control of public spaces. Consider the ‘Clean City’ law, which was enforced in 2006 in the city of São Paulo in order to remove all advertisements from public spaces, including buildings, posters, billboards, buses and taxis; even handing out pamphlets. Such a policy solution directly responds to market nudge stacking, which is exacerbated by the extension of public communication channels. It significantly alleviates the extent to which consumers are exposed to market nudges. Of course, a policy solution need not go as far as the Clean City law in creating nudge-free public spaces. We may sometimes find it more appropriate to set up specific advertisement-free city zones, since our efforts at minimizing the detriments of market nudges should be counterbalanced with bad side-effects to markets that policies may cause. But we need not think of the proposal as too radical or ruinous to market efficiency. Consider that our virtual space can already be purged of advertisements with the help of software – so-called adblocks – and that web marketers can ask users to opt out.

Some regulations will concern both content and space. A famous field experiment by Zajonc and Rajecki (1969) shows that the positive attitudes towards stimuli can be enhanced merely by repeated exposure. This finding easily explains why marketers put so much importance on public presence and flood the public space with ads. In response, public officials could introduce regulations that limit the extent to which a particular company can use public spaces for advertising.

Finally, we might believe that, sometimes at least, outright bans are justified. For instance, the European Parliament and European Council (2003) introduced a ban on all tobacco advertising in print, radio, online and in sports events. Or think of bans with respect to advertising for children, which are in place in countries such as Norway. Such strong prohibitions can be justified on autonomy grounds, e.g. when they target potentially addictive products and services, or when the targeted audience is particularly vulnerable.

6. Conclusion

The popularity of government nudges in academic and policy circles has sparked little debate about the moral status of behavioural techniques used by for-profit market agencies. This academic oversight may be somewhat understandable, but is nevertheless significant. Our paper is an attempt at righting this wrong. We show throughout that on the grounds of autonomy-related objections advanced against government nudging, market nudges turn out to be much more threatening to autonomy. They trigger heuristics in much the same way as government nudges, but in addition, they are not constrained by principles of mildness, sensitivity to the interests of nudgees, and public justifiability, and they flood the market setting as a result of businesses aiming to keep up with their competition.

The impact of market nudges on our autonomy may not seem like much – we are so used to having them around. But if we are convinced about the effectiveness of government nudges, and if we acknowledge the profound effects of advertising, then these realizations should lend credence to the threatening character of market nudges. This threat is not overstated, we argued, just because we may learn to expect attempts at manipulation in the market, or because behavioural influences are pervasive in all areas of social life. Finally, we showed that a social environment with fewer market nudges is conceivable and not overly demanding for policymakers to design. Yet, even when this is not the case, the normative foundations laid here will, we believe, outlive our current institutional limitations and guide the future policymaker.