Transactional fairness in consumer markets

Abstract

There is growing public concern about the ‘unfairness’ of many pricing practices that have become common in consumer, particularly digital, markets. Industrial and behavioural economists have developed theories that explain the conditions under which these practices are profitable for firms, and their implications for consumer welfare. We identify a mismatch between the welfare economic principles used in this theoretical work and the normative perspective in which these practices are viewed as unfair. We develop a concept of ‘transactional fairness’, grounded in the normative approach of Sugden's Community of Advantage, that is reflective of public concerns. Transactional fairness is complementary to established criteria of economic efficiency and distributional equity but is based entirely on the relationship between individual buyers and sellers. It establishes clear principles with realistic information requirements that are appropriate for compliance by firms. Regulation based on this approach can help to restore public faith in markets.

There is a growing public concern, expressed in the media, in public debate and by consumer advocacy groups, about what is thought to be the unfairness of many trading practices that have become common in consumer markets. Undoubtedly, current economic and technological trends are facilitating new forms of price discrimination and data harvesting by firms. Among these trends are the increasing importance of personalised and digital services in the economy; the shift in selling strategies from sale to rental and from payment-per-item to subscription payment; the shift to direct debit payment and auto-renewal for ongoing service contracts; and developments in information technology that give firms access to data about the individual characteristics of their customers and allow firms to use this to personalise prices and to commercialise personal data in other ways. The interactive nature of the user interfaces of websites and apps allows firms to influence consumers’ access to and processing of information in ways that are not possible in bricks-and-mortar or mail-order retailing. Many of the practices that have emerged from these trends – for example, automatic reversion to an unfavourable tariff at the end of a fixed-term energy contract, ‘back-book’ bank customers being left on unfavourable accounts that are no longer available to new customers and bundled contracts for ‘mobile phone plus usage’ that continue by default after the phone has been paid for and a SIM-only contract would be much cheaper – are widely perceived as unfair.

Legislators and regulators are conscious of this concern and are responding to it. In the USA, it contributed to the political climate in which the 2010 Dodd-Frank Wall Street Reform and Consumer Protection Act was passed, setting up the Consumer Financial Protection Bureau to regulate retail financial markets. In the UK, regulatory agencies have responded with inquiries into unfair trading practices and by introducing regulations designed to limit their use.¹ The European Union (EU) has recently standardised and codified consumer protection measures in the New Deal for EU Consumers (2020).² If economics is to remain relevant and help regulators avoid pitfalls and inconsistencies, it needs to provide criteria for assessing such interventions.

Industrial and behavioural economists have created a large body of theory to explain the conditions under which various kinds of price discrimination are profitable for firms, their implications for the prices faced by consumers, and the likely effects of different regulatory interventions. A recurring finding in this literature is that many emerging forms of price discrimination in non-monopoly markets work only because some consumers – usually, those who are discriminated against – do not behave like the rational agents of neoclassical economics. These findings have created problems for the analysis of the efficiency and distributional aspects of price discrimination. Economic efficiency is standardly defined in terms of the satisfaction of individuals’ preferences, but this benchmark is undermined if the rationale of regulation is that consumer behaviour is not always rational. One response is to argue that analysis should be based on consumers’ ‘true’ preferences, but this raises the question of how such preferences should be defined and how they can be observed. A second problem is that the distributional effects of policies are standardly assessed in relation to categorisations of individuals by income or their degree of social deprivation. However, such categories may not be the most relevant for assessing the fairness of pricing practices that discriminate between consumers according to their degrees of economic rationality.

The premise for our paper is that there is a mismatch between, on the one hand, the moral perspective in which trading practices are viewed as unfair and, on the other, the normative principles on which both neoclassical and behavioural welfare analyses are based. Those principles support a view of public policy-making as an exercise whose objective is the maximisation of social welfare. In public debate, in contrast, trading practices are often judged to be unfair by virtue of transaction-specific properties – properties of the relationship between a firm and a customer, considered without reference to wider social effects. As a result, when regulators or courts look to economics for guidance about fair pricing, they struggle to reconcile two fundamentally different normative approaches. Discussions with regulators and government economists have convinced us that this mismatch is an obstacle to the design of coherent policy responses to concerns about unfair trading practices.

Our paper is an attempt to resolve this tension by developing a concept of transactional fairness that can represent the intuitive sense of fairness that is invoked in public debate about trading practices, and that can be used by regulators and judges and in economic analysis. We present this concept as complementary to the conventional economic standards of efficiency and distributional equality and as particularly relevant for consumer protection policy. We do not claim that transactional fairness is a fundamentally new idea. It has a long pedigree in economic thought and is implicit in many aspects of consumer protection law. However, it has not been integrated into the received versions of normative economics or of the economics of regulation. Our aim is to give it an explicit formulation and to show how it can provide policy guidance.

This formulation builds on an account of market ethics offered by Sugden (2018) as part of an attempt to reconcile the ‘liberal tradition’ of economics – which is characterised by general presumptions in favour of competitive markets and against paternalism – with the findings of behavioural economics. By fitting into this scheme, our analysis of transactional fairness avoids a fault line between competition economics and behavioural economics. Traditionally, competition policy has aimed at maintaining and promoting competitive markets, in the expectation that competition will keep prices low for all consumers and incentivise firms to supply goods that satisfy consumers’ preferences. Consumers have been assumed to be rational and responsible agents, able to look after their own interests when choosing what to buy. In contrast, behavioural economics represents consumer decision-making as subject to biases and errors that can be countered by paternalistic policy interventions (e.g. Camerer et al., 2003; Sunstein & Thaler, 2003). Because transactional fairness is a procedural (rather than welfare-maximising) principle, our approach can properly recognise the empirical findings of behavioural economics without needing to invoke distinctions between consumers’ true preferences and their actual choices.

In the section ‘Two examples of “unfair” pricing’, we use two examples of allegedly unfair pricing practices to show how standard methods of normative economics – both neoclassical and behavioural – fail to take account of important aspects of fairness. In the section ‘Transactional fairness and its normative foundations’, we explain the normative foundations of our approach and present a general definition of transactional fairness. The section ‘Applying the concept of transactional fairness’ shows how this concept can guide regulatory responses to a wide range of possibly unfair trading practices. The section ‘Why transactional fairness is distinct from social welfare’ explains the fundamental differences between transactional fairness and the welfare criteria used in standard neoclassical and behavioural economics. In the concluding section, we argue that our approach can help to restore public faith in markets without deterring the emergence of new business models that provide opportunities for mutual benefit and without requiring frequent ad hoc fire-fighting interventions by regulators.

Two examples of ‘unfair’ pricing

We begin with two familiar pricing practices that illustrate mismatches between intuitive ideas about fairness and the welfare-maximising perspective taken by both neoclassical and behavioural welfare economics.

Here, and elsewhere in the paper, we follow a conventional practice in the economics literature by distinguishing between savvy and naïve consumers (‘savvies’ and ‘naïves’ for short). Savvies are rational in the sense of neoclassical theory: they act on stable, internally consistent and context-independent preferences – for short, preferences that are integrated – and on correct beliefs about relevant economic variables, including search opportunities and prevailing business practices. Naïves fall short of these standards in some way, for example, by lacking essential information, holding false beliefs or being susceptible to psychological influences that neoclassical theory assumes away.

Our first example is the practice that leads to bill shock. A firm offers a service contract with a below-cost or even zero price for its core component, along with a schedule of charges, well in excess of cost, for add-ons that consumers can incur while using the service and will be billed for afterwards. Familiar examples include late payment fees, bank charges for unarranged overdrafts and high unit prices for mobile phone usage above some threshold. Savvies avoid the add-ons and benefit from the low headline prices. Naïves incur the add-ons, either by signing up for contracts without recognising the significance of the small print or by inattention afterwards.

Our second example is the loyalty penalty. Firms offer service contracts that are subject to periodic renewal; at each renewal, the firm quotes a new price. Firms attract new customers by offering low initial prices, which are increased at each renewal by proportionately more than the increase in the firm's costs (‘price walking’). Savvies notice price increases and switch frequently between suppliers. Naïves do not notice and are penalised for their ‘loyalty’ to their original supplier.

In these examples, price discrimination is based on differences in consumers’ information about, understanding of or attention to firms’ pricing strategies. Because of this, the textbook result that price discrimination cannot occur in a competitive market does not apply. In a market with no fixed costs or barriers to entry, competition between firms can induce an equilibrium in which firms operate at minimum cost and earn only normal profit, but there is still price discrimination between naïves and savvies (Gabaix & Laibson, 2006; Armstrong & Vickers, 2012; Armstrong, 2015; Grubb, 2015).

Such pricing practices are seen by many consumers as unfair. For example, a large-scale telephone survey of UK motor and home insurance customers, commissioned by the Financial Conduct Authority (FCA), asked respondents to say whether the following was fair or unfair: ‘Mr Smith has been with the same insurance firm for 5 years and pays £500 for his building's insurance. Mr Jones, whose house is identical, asks Mr Smith's insurance provider for a quotation, and is quoted £300 for the same policy’. Only 9% of respondents thought this was ‘fair’ (FCA, 2019b, Annex 4).³ Intuitively, it is easy to understand how these kinds of price discrimination can be seen as unfair, but it is more difficult to represent their unfairness within the theoretical framework of welfare economics.

Take the case of bill shock in a competitive market with no fixed costs. In equilibrium, firms produce at minimum average cost and earn only normal profit. The profits that firms earn from naïves are competed away in the loss-leading headline prices by which firms compete to attract customers and from which savvies benefit. Suppose for simplicity that the incremental cost to a firm of supplying the relevant add-on is zero, that naïves do not anticipate the possibility of incurring this add-on, and that attentive customers can evade the add-on at no cost to themselves. Under these conditions, both savvies and naïves have the ex ante perception that they are buying only the core service and paying only the headline price; ex post, it is as if the firm's loss from selling below cost is made up by an arbitrary levy on naïves. Apart from the possible inefficiency that results from overconsumption of the under-priced core service by both types of consumers – an effect that seems orthogonal to concerns about fairness – the effect of price discrimination is a pure transfer from naïves to savvies. In standard economic analysis, an evaluation of this effect would be a distributional judgement based on the relative income of savvies and naïves. If naïves are predominantly cash-rich, time-poor consumers who can afford to be inattentive to add-on prices, that transfer might be judged as an increase in welfare, but many would still consider that the naïves are being treated unfairly. A further source of unfairness arises when the user interface makes it more difficult for customers to avoid the add-on or to buy from a different supplier. Once again, the main welfare effect may be the transfer from naïves to savvies, which does not necessarily capture the sense of unfairness.

Next, take the case of the loyalty penalty. The unfairness of these practices is similarly difficult to express in a welfare-based analysis: loyal customers need not be disproportionately poor, and the loyalty penalty may even be welfare-enhancing. To see this, note that price competition can be effective only if consumers seek out low prices, and consumers would have no reason to compare firms’ offers if they already knew that all offers were exactly the same. In realistic models of price competition, some price dispersion persists in equilibrium. If firms can price discriminate, they have the incentive to set a low price for the savvies who search and a high price for the naïves who don't, for example, by offering the low price in a complex tariff that only savvies can decipher. Under some conditions, this reduces the average price paid and raises average welfare (Armstrong, 2015).⁴ Nevertheless, inasmuch as the firm is ‘hiding’ offers for which its naïve customers are eligible, one might still agree with the survey respondents who thought that Mr Smith's insurance provider had treated him unfairly.

Even if naïves are distinguished by their cognitive vulnerability, for example, due to age, infirmity or lack of financial understanding, it is not obvious that effects on them should be given greater distributional weight than effects on savvies. Suppose that Arthur is an 85-year-old widower with a good index-linked pension but declining mental powers and that Bella is a 30-year-old low-income single mother who takes care when using her bank account. As a result of his poor memory, Arthur incurs high add-on bank charges from which Bella ultimately benefits. Viewed from a consequentialist perspective, the effect is a small transfer of income from someone for whom the marginal utility of income is relatively low to someone for whom it is higher. But one might still think that, by taking advantage of his vulnerability, Arthur's bank has treated him unfairly.

Why do these examples pose problems for economics? If economics has its own coherent methods of normative analysis, and if, according to an analysis conducted using those methods, there is nothing objectionable about some pricing strategy that the general public regards as unfair, isn't the problem simply that the general public has an insufficient understanding of economics? We believe that such a response would be inadequate for (at least) the following reasons.

First, much work in economics is framed as advice to market regulators, courts, and economic policy-makers. In a democratic society, regulators, judges and politicians are ultimately constrained by, and so need to take account of, citizens’ judgements about economic matters. If regulation or court decisions will, in fact, be influenced by citizens’ conceptions of fairness, those conceptions need to be codified so that decisions can be anticipated and effective in modifying ex ante behaviour. Market regulation has to be more than an unsystematic collection of case-by-case ex post decisions. It needs to be based on general principles that are stated publicly and are commonly understood by actors in the economy.

Second, if a market economy is to be politically sustainable, consumers must have a general sense that the market treats them fairly. If a firm is able to use pricing practices that are generally perceived as unfair, this is liable to undermine trust, not only in that particular firm but also more generally in markets, regulators and the market system.⁵ Effects of this kind are particularly dangerous in markets such as financial services, where trust is fundamental to the nature of the product.

Our third reason goes deeper. The normative assumptions used in current economic analysis are not immutable. Philosophically and methodologically, they reflect the influences of utilitarianism and rational choice theory on modern welfare economics. They support a view of economic institutions as mechanisms through which unintended social outcomes are generated by the interactions of rationally self-interested individuals and a view of public policy-making as a problem of mechanism design whose objective is the maximisation of social welfare. But, as we explain in the next section, other ways of understanding economic life have equally deep intellectual roots.

Transactional fairness and its normative foundations

Our concept of transactional fairness is grounded in an approach to normative economics that is significantly different from neoclassical and behavioural welfare economics but has a long intellectual history. Adam Smith's (1776/1976: 456) metaphor of the ‘invisible hand’ is often (and not unreasonably) cited as a precursor of the First Fundamental Theorem of welfare economics. But in another famous passage, Smith characterises the market in terms of procedural properties – as ‘the obvious and simple system of natural liberty’ in which every man ‘is left perfectly free to pursue his own interest in his own way’ (p. 687). This is a view of the market as a system in which individuals have open-ended opportunities for voluntary transactions with one another. In a similar spirit, transactional fairness is based on principles of opportunity rather than welfare. In one of the most influential philosophical works of the 20th century, Rawls (1971) proposed a procedural account of ‘justice as fairness’ based on a conception of a well-ordered society as a ‘cooperative venture for mutual advantage’ (p. 4). Our analysis builds on the related idea of the market as a network of cooperative interactions directed at mutual benefit.

This normative approach has been developed by Sugden and co-authors to take into account the empirical findings of behavioural economics (Sugden, 2004; Bruni & Sugden, 2013; Infante et al., 2016; Sugden, 2018). This research programme focuses on individuals’ opportunities, rather than on how far, given those opportunities, their preferences are satisfied by the choices they make. One of its results is that a competitive equilibrium of a market economy can be defined without assuming that individuals have integrated preferences. In such an equilibrium, every group of individuals has the collective opportunity to make any feasible transaction among themselves that they might conceivably find mutually acceptable (Sugden, 2004; McQuillin & Sugden, 2012).

Why should the market's tendency to create opportunities be viewed as desirable, even if individuals do not act on integrated preferences? The key ethical idea can be expressed as the following.

Principle of Voluntary Market Transactions. A well-functioning market is a domain in which individuals engage voluntarily in cooperative interactions.

For a transaction to be voluntary, it must take place only with the consent of all participants and without hindrance to potential transactions that those participants might have chosen instead.

Within a voluntary interaction, behaving ethically is playing one's fair part in a cooperative scheme. Sugden (2018: 256–281) expresses this ethic of cooperation in the following principle.

Principle of Mutual Benefit. When participating with others in a voluntary interaction, and for as long as others’ behaviour in that interaction is consistent with this very principle, behave in such a way that the other participants are able to satisfy normal expectations about the consequences of the interaction for them.

Normal expectations about a given type of interaction are beliefs that (a) are at least approximately correct as a description of actual behaviour in that type of interaction, (b) most members of the relevant population can reasonably infer from their own experience and information and (c) are, in fact, held by most members of that population.⁶

The Principle of Mutual Benefit applies to voluntary interactions in general, but market transactions are a paradigm case.⁷ If some class of interactions is voluntary, and if there are normal expectations about how people behave within such interactions (the practice for that class of interactions), then your choosing to participate is a signal that conditional on the others also choosing to participate: (a) you intend to conform to the practice; (b) you expect the other participants to conform; (c) you expect to benefit from the interaction; and (d) you believe that the others expect to benefit. By virtue of these properties of participation decisions, the existence of a practice is an opportunity for mutual benefit. By conforming to a practice in a specific interaction, you play your part in a cooperative activity that involves you and the other participants. At the same time, you are also playing your part in a wider scheme of cooperation: by conforming to a practice, you reinforce the common expectation that it will be followed, and so help to sustain opportunities for others to achieve mutual benefit by following it.

The essential idea behind our concept of Transactional Fairness is that market transactions are fair if the participants adhere to the Principle of Mutual Benefit and, in accordance with the Principle of Voluntary Market Transactions, do not obstruct one another's opportunities to transact with others. A transaction is an economic interaction between specific participants, each of whom has chosen to take part in it. Considering any given transaction, the question that a concept of transactional fairness has to answer is: ‘In this transaction, is each participant treating each other participant fairly?’ Notice that this is a property of individual transactions and not, as efficiency and income distribution are, of an economic system as a whole. In this paper, we focus on the fairness of firms in their dealings with consumers.⁸

In the business-to-consumer context, and drawing on the above definition of ‘normal expectations’, our definition of transactional fairness is as follows:

Transactional Fairness. Transactional fairness requires that a firm acts in such a way that consumers with normal expectations about consumer offers in the relevant market are able to understand the consequences of transacting with that firm (No Deception) and are not hindered from terminating a relationship with the firm or from transacting with alternative sellers (No Hindrance). If the transaction involves a consumer with observably impaired capacity to understand the consequences or to search for alternative offers, fairness further requires that the firm does not knowingly exploit this (No Exploitation of the Observably Vulnerable).

Applying the concept of transactional fairness

We now explain the components of our definition of transactional fairness in more detail and discuss their application to particular examples.

Normal expectations

If a market economy is to be open to the development of new business models, standards of transactional fairness cannot be entirely context-independent. For example, one might reasonably claim that the standards of fairness that apply to transactions between buyers and sellers at a car-boot sale or flea market are laxer than those that apply to transactions between a department store and its customers. This need not be problematic if, in each type of market, consumers know what to expect and enter each transaction voluntarily. Our concept of normal expectations allows us to represent this idea.

Normal expectations about pricing practices can be interpreted as default settings for the terms of implicit contracts between firms and consumers. Implicit contracts, based on (often market-specific) expectations as to what are acceptable commercial practices, are common in business-to-business transactions.⁹ These norms reduce transaction costs and facilitate flexible adaptation to changing circumstances. They can be enforced by reputation, withdrawal of future business and possibly resort to contract law.¹⁰ Implicit contracts between firms and consumers are similar in principle but often less enforceable because each consumer typically accounts for only a small share of the firm's sales.

Transactional fairness requires each party to an implicit contract to comply with its terms. If a firm knows that different groups of consumers are likely to have different expectations, it should make public any business practices that it intends to follow and that are obviously relevant to the participation decision. If a firm relies on (and does not try to correct) customers’ mistaken beliefs about its business practices, it is engaging in passive deception. Mistaken beliefs are particularly likely in periods when business practices are changing rapidly, as is currently occurring through the growth of digital markets. Many of the practices that we will classify as transactionally unfair are particularly prevalent in digital markets, where they exploit expectations that were normal in bricks-and-mortar retailing.

No Deception

Deception, either active or passive, undermines the intention of mutual benefit that allows a transaction to be understood as cooperative. There is deception, and therefore transactional unfairness, if a consumer is enticed into a transaction by a firm's use of information that is misleading, or by its hiding obviously relevant information about the transaction. ‘Hiding’ includes presenting information with misleading salience that draws attention away from what is likely to be most relevant. The prohibition of misleading practices is familiar territory in consumer protection, but the absence of an agreed normative foundation currently makes consumer law slippery to apply. The concept of transactional fairness can provide such a foundation.

Deception operates relative to normal expectations. ‘Obviously relevant information’ includes information about the prices of add-ons that are effectively unavoidable (e.g. delivery charges, fees for debit card payments for online sales) or that most customers would expect to buy in combination with the main product (e.g. product-specific ink cartridges for printers).

Where the normal expectation is for a consumer to be offered available prices, it is unfair for a firm to hide information about other relevant tariffs that it offers and for which the consumer is eligible. This rules out forms of price discrimination that rely on consumers’ lack of information about the firm's own prices.¹¹ For example, it would be unfair for a train operating company to sell ‘any time’ tickets to walk-up customers in off-peak periods without informing them about its cheaper off-peak tickets. If, instead of quoting take-it-or-leave-it prices, the firm is stating an initial offer that it is willing to negotiate, or that it is willing to reconsider if a potential customer can show that a rival has quoted a lower price, it should make this clear as a property of that offer. Similarly, it is unfair if the firm secretly personalises offers to consumers when the normal expectation is that prices are fixed. Nevertheless, as we make clear in the section ‘Price discrimination is not intrinsically unfair’, price discrimination is not inherently unfair as long as it is not deceptive.

As we noted in the introduction, there is an increasing tendency for products to be sold on indefinite or default contracts, for rental deals to replace sales of durable goods and for subscription contracts to replace payment-per-item selling. Such business models involve an ongoing relationship between the firm and ‘its’ customer (or ‘client’). This creates a continuing implicit contract which expands the scope of the No Deception condition.

That condition implies that there is transactional unfairness if a firm attempts to retain an existing consumer by giving misleading information, or by hiding obviously relevant information about the terms of the continuing transaction. Thus, it is unfair if the firm does not periodically provide an existing customer with relevant information about changes in its prices, or about the absence of a price change for a product whose cost of supply is falling over time. Several UK regulators have recently adopted policies requiring firms to provide such information when contracts are renewed (e.g. the requirement that insurance renewal documents report the customer's previous price alongside the renewal quote).

As in the case of pre-purchase fairness, it is unfair for a firm to hide information about alternative tariffs that it offers and for which the customer is eligible or about its willingness to renegotiate initial renewal quotes.¹² While these principles limit many forms of price discrimination between newly acquired and long-standing customers, they do not rule out discrimination that is mutually beneficial. For example, low-price introductory offers (exclusive to new customers) are compatible with intentions for mutual benefit if the low price is paid only for an introductory period, and if the intention is to allow new customers to sample the firm's product or to compensate them for costs of searching and switching.

There is a growing tendency for firms to use personalised pricing for products that are sold on continuing contracts. Rather than offering a publicly displayed range of tariffs from which consumers can choose, subject to specified eligibility criteria (e.g. new customers only), a firm determines the offers it makes to individual consumers on the basis of its information about their particular circumstances or previous purchases. Personalised pricing is particularly prevalent in insurance markets, where normal underwriting practice requires customers to provide information that identifies risk-relevant personal characteristics, and the price of a given level of cover varies according to those characteristics. It is a small technical step to extend this practice to ‘margin optimisation’, that is, varying prices according to the profitability of different classes of customers when profitability is not related to risk. For example, in making price offers to new customers, insurance firms routinely take account of differences in consumers’ propensities to renew their contracts and buy add-ons (FCA, 2019b).¹³ If there is a normal expectation that personalisation is risk-based, such undeclared margin optimisation is a form of passive deception.

So far, we have considered deceptive pricing practices, but there can also be deception about non-price properties of firms’ offers. For example, consider an online retail platform that sells its own products alongside those of other producers, or an online booking site that takes commission from suppliers. If such an intermediary gives greater prominence to (or makes implicit recommendations in favour of) its own products or those of suppliers who pay higher rates of commission, failure to disclose this practice is a form of passive deception. Another potential source of passive deception arises when a core product is cross-subsidised and the quid pro quo for the transaction is not obvious to the consumer. For example, search or social networking services are typically offered at a zero price, with profits made by using the consumer's data to sell targeted advertising space or monetised in some other way. There is passive deception if the consumer is not made aware of possible negative consequences of the transaction for her (e.g. junk mail, invasion of privacy or personalised prices) before agreeing to it.

One might ask whether fairness requires that a firm gives potential customers’ relevant information not only about its own offers but also about its competitors’ offers.¹⁴ If consumers can generally be expected to be aware of the existence of alternative suppliers, not revealing a rival's price is not a violation of transactional fairness: the information that is not being revealed is not information about the transaction whose fairness is being assessed. More fundamentally, our definition of transactional fairness is grounded in an ethical conception of a well-functioning market economy. The practice of firms revealing their competitors’ prices could not persist as a normal expectation in a competitive market. Such a practice would undermine both consumer responsibility to search and rivalry between firms – mechanisms that are fundamental to the working of the market.

No Hindrance

It is fundamental to our conception of transactional fairness that the parties to a market transaction should interact voluntarily. The No Hindrance condition upholds this principle by requiring that consumers are not hindered from terminating a relationship with a firm or from transacting with another firm. The first clause is in the spirit of consumer law; the second is in the spirit of competition law.

It is transactionally unfair if a firm uses pricing practices designed to deter consumers from searching for competitors’ offers. For example, industrial economists and regulators have analysed firms’ use of time-limited (or exploding) offers – offers that must be accepted or rejected within a time frame that is too short to allow a potential buyer to search for other offers. Such practices inhibit voluntary choice. By creating disincentives to search, these practices also tend to raise prices (e.g. Office of Fair Trading, 2010; Armstrong & Zhou, 2016), but their transactional unfairness is independent of that tendency.

In the context of ongoing relationships between firms and customers, there is a restriction on a customer's opportunities to transact if the firm makes it difficult for her to cancel a contract for which renewal is the default. Familiar hindrances to cancellation include procedures that require consumers who signed up online to cancel by mail or phone (sometimes using phone lines with very slow answering services) and online interfaces in which cancellation options are not easily visible or involve unnecessary sequences of operations. A related practice, now prohibited by the relevant UK regulator, is for operators of mobile phone networks to sell devices that are ‘locked’ to the provider's network in ways that many customers find difficult to undo (Ofcom, 2020: par 1.6–1.8). A useful benchmark from which to assess hindrance is exit/entry equivalence – that consumers find it as easy to exit a contract as it was to enter.¹⁵ Unfair hindrance applies to indefinite contracts and those with default renewal; it does not release consumers from fixed-term contracts that were fairly entered into (e.g. fixed-term mortgages with pre-specified interest rates).

Analogously with passive deception, there can be unfair passive hindrance. For example, if a contract is subject to periodic renewal by direct debit, it is unreasonable to expect consumers to remember renewal dates in the absence of reminders or renewal statements. Fairness requires a firm to give customers sufficient notification of upcoming renewal dates to allow them to search for alternative offers.

Notice that the No Hindrance condition does not require firms to make public offers. This leaves room for firms to use practices that, at the industry level, impose barriers to search. Personalised pricing is an example. Even when there is common knowledge that firms are using this practice, personalised pricing makes the terms on which a firm trades with each customer private to those two parties. To the extent that this makes searching for the best offer more onerous for consumers than comparing publicly announced prices or tariffs, personalised pricing may have anti-competitive effects that regulators wish to take into account.¹⁶ But a concept of fairness within an individual transaction should not require that the participants reveal to others the terms on which they are trading.

No exploitation of the observably vulnerable

Our No Deception and No Hindrance conditions operationalise the Principles of Mutual Benefit and Voluntary Market Transactions, which we argue are fundamental to an ethical market economy. However, the ethical force of those principles derives from a conception of individuals as competent choosers.

We define a consumer as being transactionally vulnerable in relation to a given transaction if her capacity to understand the consequences of that transaction or her capacity to search for alternative offers is impaired. In such cases, the voluntariness of market transactions is compromised. Fairness in a transaction with a vulnerable consumer requires a third condition – No Exploitation of the Observably Vulnerable. Unlike No Deception and No Hindrance, this condition involves benchmarking across transactions.

A transaction with a vulnerable consumer is unfair if the firm either targets or responds to an observably vulnerable consumer with an offer that is on worse terms than is generally available to the non-vulnerable. ‘Observability’ includes both direct observation and statistical inference. ‘Impairment’ may be due to long-lasting cognitive limitations (e.g. young children and individuals with dementia), temporary distress (e.g. a recently bereaved person planning a funeral) or inability to access commonly used sources of information (e.g. lack of internet access may be relevant for some products). For firms that sell to both vulnerable and non-vulnerable consumers, fairness when transacting with the vulnerable requires that, if price discrimination is part of the business model, it does not exploit transactional vulnerability.¹⁷ For firms whose customers are predominantly vulnerable, fairness may require regulation to determine an appropriate benchmark.

Our definition of transactional vulnerability is deliberately narrower than many uses of the term ‘vulnerable’ in current UK discussions about market regulation. For example, a recent report by the FCA identifies ‘protecting vulnerable consumers’ as a ‘key priority’ and uses a definition that classifies 50% of British adults as ‘potentially vulnerable’ (FCA, 2019a: par. 1.1). A vulnerable consumer is defined as one ‘who, due to their personal circumstances, is especially susceptible to detriment, particularly when a firm is not acting with appropriate levels of care’ (par. 2.1).

This definition goes beyond transactional vulnerability in two ways. First, it conflates impaired decision-making capacity, which affects what counts as fair in a transaction, and factors such as low income, low savings, indebtedness and job insecurity, which raise the stakes of making good decisions and so affect how much a consumer would be harmed if she were treated unfairly. The degree of harm caused by an unfair practice is relevant for prioritising which cases a regulator should pursue, but it should not be part of the definition of unfairness itself.

Second, the FCA report treats being ‘more prone to certain behavioural biases that negatively impact their decision making’ as a characteristic of vulnerability (par. 2.11). The implicit suggestion is that consumers whose behaviour is inconsistent with neoclassical theories of rational choice are in special need of protection. Given that this is a normal aspect of human psychology and that such inconsistencies can be found across the whole domain of everyday consumer decision-making, this casts a very wide net. Indeed, if accepted, it would undermine the normative value of consumer choice in a market economy. As we will explain in the section ‘No assumptions about latent preferences’, transactional fairness does not depend on an unrealistic view of human psychology or try to distinguish between true preferences and biases.

Why transactional fairness is distinct from social welfare

As we made clear in the introduction, we are not presenting transactional fairness as a standard that can replace the conventional economic standards of efficiency and distributional equality. We claim only that it is a significant and distinct normative concept. In this section, we explain how it differs from the normative standards used in neoclassical and behavioural welfare economics.

Action-guiding for whom?

Transactional fairness is concerned with ethical standards that can be action-guiding for individual firms and consumers. In this respect, it is unlike efficiency- and welfare-based approaches to normative economics, which are intended to guide the actions of regulators or policy-makers.

Because efficiency and social welfare are properties of a whole economy, individual firms and consumers cannot be expected to base their decisions directly on such criteria. In contrast, principles of transactional fairness are formulated in terms of information that is available to the firms and consumers who are expected to be guided by them. This condition imposes limits on how far transactional fairness can require firms to meet their customers’ preferences. In many retail settings, firms cannot know the preferences of particular customers.¹⁸ However, a firm can be expected to know that consumers generally prefer lower to higher prices and to be aware of consumers’ normal expectations about business practices in the markets in which it operates.

If the principles of transactional fairness are accepted as action-guiding ethical standards for firms, it is reasonable to expect firms to have taken account of them when choosing their trading practices. As an aid to compliance with transactional fairness, we propose the following complementary requirement of self-assessment by firms:

Challenge of Public Explanation. A firm should be able to defend the rationale for its consumer trading practices, to locate them within a business model based on mutual benefit between it and its customers, and be willing to present that rationale publicly.

By the ‘rationale’ for a trading practice, we mean a genuine account of the firm's reasons for choosing to use it. By ‘mutual benefit’, we mean that the firm can pursue the interests of its owners within the constraints of No Deception and No Hindrance. To be clear, we are not proposing that the fairness of specific practices should be judged by public opinion, but only that a firm's compliance with a well-defined standard of fairness can be publicly verified.

An important part of the rationale of a trading practice is the identification of the consumer groups with whom the firm expects to trade. An implication of mutual benefit is that: ‘Products and services marketed and sold in the retail market are designed to meet the needs of identified consumer groups and are targeted accordingly’. This wording is part of a guideline formulated by the FCA (2019a: Annex 1, par 17). Similar guidance could be extended naturally to tariffs that are offered to retail consumers.

In a less direct sense, the principles of transactional fairness are action-guiding for regulators. In this context, the role of regulation is to prohibit business practices that are clearly unfair, to enforce against unfair practices (whether or not they are explicitly prohibited) and to support public understanding of transactional fairness, for example, by demanding public explanation (with suitable evidence) of a business practice that is prima facie unfair.

Voluntariness

The principles of transactional fairness relate to behaviour within interactions that are entered voluntarily. They have nothing to say about whether an individual or firm should or should not enter into any particular transaction. It follows that these principles do not require firms to enter transactions that they do not expect to be profitable.

In this respect, transactional fairness is different from most of the concepts of pro-sociality that are represented in social preference theory. In such models, individuals are represented as having preferences for such things as benefiting other people (e.g. Becker, 1974), reducing inequality between themselves and others (e.g. Fehr & Schmidt, 1999; Bolton & Ockenfels, 2000) or confirming other people's expectations of benefit (e.g. Battigalli & Dufwenberg, 2007). In all these models, pro-sociality is represented as a willingness to make some form of self-sacrifice. In contrast, transactional fairness is concerned with interactions that are directed at mutual benefit.

We recognise that there are exceptional situations in which a firm might be judged to have a moral obligation to engage in loss-making transactions. For example, one might think that an airline or train operating company ought to carry disabled passengers at its normal fares while providing them with the additional assistance they need, or that it ought to waive cost-recovering rebooking charges for customers who need to change their plans because of family emergencies. But these are best understood as obligations of humanity or decency, not of transactional fairness.

Irrelevance of externalities

Because transactional fairness is a property of the relationship between the participants to a transaction, external effects of that transaction on non-participants have no bearing on its fairness (other than in contributing to normal expectations). Thus, in our example of bill shock, the benefit that Bella derives from Arthur's add-on bank charges is not relevant for an assessment of the fairness of the transaction between Arthur and his bank. Similarly, it might be a fact that, by increasing incentives to search, price walking tends to reduce the overall level of prices in the market for home insurance; but that fact would not be relevant to an assessment of the fairness of the transaction between Mr Smith and his insurance provider.

Irrelevance of the distribution of cooperative surplus

It might seem natural to think that a concept of transactional fairness should include principles about fairness in the distribution of the cooperative surplus, i.e., the gains from trade attributable to the relevant transaction. However, such distributional requirements are not compatible with the workings of a market economy in which firms set prices at which consumers choose whether or not to buy.

Unless prices are personalised, there will always be some marginal consumers for whom the benefit of the good they buy is almost exactly equal to the price and whose share of the cooperative surplus is therefore almost zero. In some market structures, this may be true for almost all consumers. For example, consider the market for some service that is supplied by many small firms that differ only in terms of minor product differentiation; consumers are almost indifferent between alternative suppliers. Suppose these firms have significant fixed costs but produce at zero marginal cost. In monopolistic competitive equilibrium, all charge the same average-cost price p. In any transaction between a firm and an individual consumer, the firm gains p, but, relative to her outside option (buying from another firm), the consumer gains almost nothing. Thus, the firm appropriates almost all the cooperative surplus.¹⁹ One might conclude that this market structure is economically inefficient, but it would be wrong to claim that the individual firm is treating the individual consumer unfairly.

What if a firm has market power and is able to charge a monopoly price for some discrete product? Such a firm earns a positive profit (i.e. price minus marginal cost) from every transaction. In transactions with marginal consumers, this profit appropriates almost the whole of the cooperative surplus, but for consumers whose reservation price is particularly high, the firm's proportionate share of the surplus may be quite low. Applying the principle of fair sharing of surplus, one might conclude that the firm was treating marginal consumers unfairly, but the same principle would imply that consumers with high reservation prices were treating the firm unfairly. Again, there is a mismatch with ordinary ideas about fairness.

Of course, there are transactionally unfair ways of appropriating surplus – namely, by deception and hindrance – and these may be more accessible to firms with greater market power. Nevertheless, transactional unfairness is located in the deception or hindrance, not in the market power itself.

Price discrimination is not intrinsically unfair

In the section ‘Applying the concept of transactional fairness’, we categorised various forms of price discrimination as transactionally unfair. Many of those pricing practices discriminate against naïve consumers and in favour of savvies. In our analysis, however, the unfairness of such practices derives from properties of deception or hindrance that are located in the transaction between the firm and the naïve consumer, and not in the difference between the firm's treatment of the two classes of consumers. This is fundamental to our concept of transactional fairness as a property of individual transactions.

Recall that we are presenting transactional fairness as complementary to the standard criteria of efficiency and distributional equality. In terms of each of the latter criteria, price discrimination by profit-seeking firms can have both positive and negative effects. These effects can properly be matters of concern for regulators, but they are outside the scope of transactional fairness.

We note that price discrimination can also be unfair in ways that are not transactional and that the standard economic criteria of efficiency and distribution do not take into account. That people should be treated equally, irrespective of ‘protected’ characteristics such as age, disability, sex, ethnicity, religion and sexual orientation, is a very widely held ethical principle, upheld by law in most democracies. However, this does not mean that pricing practices that discriminate on the basis of protected characteristics are thereby transactionally unfair. We believe that clarity is best served by treating transactional fairness as a distinct form of fairness.

Non-paternalism

Transactional fairness is a non-paternalistic principle, grounded in a conception of voluntariness. It requires that the offers that firms make to consumers are presented without deception and without obstructing consumers’ access to rival offers, but it is up to each consumer to decide whether to accept any particular offer. Within these constraints, transactional fairness does not require firms to make their own judgements about a consumer's welfare.

In contrast, a welfare-based approach to normative economics might reasonably recommend paternalistic policy interventions in situations in which consumers are judged to be acting contrary to their best interests. It might also – as some behavioural economists indeed do – address firms directly and recommend them to act paternalistically towards their customers.²⁰ For example, a frequent claim in behavioural economics is that, because of failures of self-control, consumers are liable to choose superficially attractive products, contrary to their true preferences or self-acknowledged long-term interests. This hypothesis is often represented in dual-self models in which a person's rational self (the ‘Planner’) can be subverted by an impulsive self (the ‘Doer’). Thaler and Sunstein (2008: 41–49) endorse this model and recommend ‘choice architecture’ that ‘nudges’ consumers towards the options that their Planners would choose. Thaler (2018) distinguishes between ‘conscientious’ choice architects who ‘nudge for good’ and firms whose marketing nudges ‘encourage buyers in order to maximize profits rather than to improve the buyers’ welfare’ and proposes a ‘sludge clean-up campaign’ to eliminate the latter kind of nudge.

Taking a line similar to Thaler's campaign against sludge, Oliver (2023: 134–148) argues for a type of regulation he calls ‘budging’. A budge is ‘a regulation against the negative externality of a behavioural-informed manipulation’ by a firm (p. 144). A firm engages in such manipulation if it uses psychological cues (‘the behavioural influences’) to induce its customers to buy ‘more of [its] products than they would otherwise desire’; this constitutes ‘an infringement upon the notion of a free and fair exchange’. Treating the manipulation-induced change in purchases as an externality of the transaction and as a harm imposed on the consumer by the firm, Oliver argues that budges are not paternalistic and are thereby compatible with the classical liberalism he espouses (pp. 137–141).

Thaler does not spell out his conception of ‘conscientious’ marketing, and Oliver does not define the notion of ‘free and fair exchange’. But these categories would seem to exclude any non-informative advertising or packaging that cues associations of ideas between a product and pleasurable sights or experiences.²¹ In distinguishing between nudge and sludge, Thaler seems to be treating profit-maximisation as a morally questionable objective for a firm. Transactional fairness takes a different perspective. In choosing which (non-deceptive and non-hindering) offers to make and how to present those offers in the most attractive light, it is not unfair for a firm to seek to maximise profit.

We think it is important to maintain a clear distinction between paternalistic and non-paternalistic justifications for regulatory policies. We are not asserting that regulators should never be paternalistic, but the legitimacy of a paternalistic intervention is best achieved by some form of democratic mandate, whereas a non-paternalistic intervention can usually be delegated to an independent regulator or the courts.

No assumptions about latent preferences

Some readers may be thinking that the argument in the previous section overlooked the effects of psychological biases. Many of the empirical findings of behavioural economics are routinely described as ‘biases’ that interpose between individuals’ true preferences and their choices. In the literature on user interface design, ‘dark patterns’ (or ‘deceptive design patterns’) are often defined in a way that includes interfaces that ‘exploit cognitive biases’ (e.g. Mathur et al., 2019: 812). This terminology might suggest that a firm should be deemed to act unfairly if it knowingly activates psychological biases in a way that induces its customers to act contrary to their true preferences. Such a principle of fairness would be in accord with a normative approach that is now standardly used by behavioural economists, but which we argue is flawed.

Behavioural welfare economics was first proposed in two influential manifestos, by Camerer et al. (2003) as ‘asymmetric paternalism’ and by Sunstein and Thaler (2003) as ‘libertarian paternalism’, and subsequently developed by (for example) Kőszegi and Rabin (2007) and Bershears et al. (2008). As in neoclassical welfare economics, the normative criterion is the welfare of individuals, defined in terms of preference satisfaction. The difference is that behavioural welfare economics does not assume that preferences are reliably revealed in choice, even when individuals have full information about the relevant market. Instead, each individual is implicitly assumed to have latent (or ‘true’) preferences which would be revealed in her decisions if she had complete information and was not subject to biases or errors attributable to deficiencies of cognitive capacity, attention or self-control.

However, the claim that naïve choices can be explained by the interaction of latent preferences and psychological biases is questionable. Infante et al. (2016) argue that the concept of latent preference lacks psychological foundations and, in consequence, is explanatorily redundant.²² How far an individual's decisions are context-dependent is an empirical question, but which contexts induce true preferences is not. The case of guaranteed asset protection (GAP) insurance provides an illustration. GAP insurance is often sold as an optional add-on when a new car is bought; typically, an equivalent product could be bought as a stand-alone purchase at a significantly lower price (FCA, 2018a). The evidence suggests that, for many consumers, willingness to pay for GAP insurance is context-dependent: although the product is more expensive as an add-on than as a stand-alone purchase, over 90% of GAP sales are add-ons (FCA, 2018a).²³ Is psychological bias in the saleroom (over-attention to the enticing features of the new car, susceptibility to saleroom pressure) defeating a true preference for not insuring, or is psychological bias afterwards (lack of engagement with financial matters, procrastination) defeating a true preference for insurance? To pose such a question presupposes a benchmark of correct preferences or, equivalently, of the correct distribution of attention between different elements of a decision problem. These are benchmarks that economics does not supply and psychology does not need. We conclude that, except in the most uncontroversial cases (e.g. assuming that consumers prefer to buy given goods at lower prices), judgements about transactional fairness should not depend on assumptions about latent preferences.

Conclusion

Our aim in this paper has been to contribute to the development of general principles by which the fairness of firms’ trading practices can be assessed. At a fundamental level, the problem we have addressed is that of maintaining clear and consistent expectations in markets in which practices evolve over time. Currently, an undesirable dynamic seems to be at work. Firms discover new strategies, perhaps made possible by advances in technology, for making profit within the constraints imposed by existing regulations. These strategies can drift into becoming common practices in a market, viewed by firms as ethically acceptable, even though they involve passive deception or hindrance of consumers who are not aware of how business practices have changed. It is only after sufficient evidence of unfairness has accumulated that interventions or new regulations are introduced. This persistent misalignment of expectations, combined with the perception of regulation as ad hoc fire-fighting, undermines public trust in the market system.

We hope that our paper will be useful to regulators who have to grapple with these difficult practical problems. But it should also be seen as a contribution to normative economics, showing that there is an economically significant concept of fairness – transactional fairness – that does not reduce to the familiar concepts of welfare, efficiency and distributional equality. Unless this form of fairness is recognised, the purpose of regulation cannot be fully understood.

Our perspective is that a fundamental role of a market regulator should be to uphold a coherent ethical conception of a well-functioning market as a network of mutually beneficial cooperative interactions. Viewed from the perspective of firms, such a market is a space in which each firm is free to seek profit by proposing transactions to consumers, provided it does so without deception and without hindering transactions in which it is not involved. Viewed from the perspective of consumers, such a market offers an array of opportunities for transactions that each individual is free to enter voluntarily, knowing what to expect, and remaining free to leave.