Law and Economics of the Withdrawal Right in EU Consumer Law

2.1 Personalised Mandatory Rules

Nowadays, there is a general trend toward personalising private law. This idea has been initially advanced by eminent U.S. legal scholars such as Ben-Shahar and Porat (2016, 2019, 2021), Casey and Niblett (2017), Porat and Strahilevitz (2014), Sunstein (2013), and Verstein (2019), but has also found proponents in European academia – such as Busch (2019), Hacker (2017), and Luzak (2021).^[7] These scholars claim, by and large, that private law has much to gain from tailoring its regulatory apparatus to the needs of individual legal subjects (Hacker 2017). Therefore, there is an attempt to personalise private law across different regulatory tools, such as disclosures, defaults, and mandates. With the help of personalisation – that is, by considering factors such as the degree of rationality or willpower, the income or wealth of each contracting party – the legal categories constructed by modern law (such as the category of consumers) could be rendered “more precise, granular and refined” to match real needs and characteristics as closely as possible (Hacker 2017, esp. pp. 659, 676). Of course, the holistic attempt to personalise private law raises significant privacy and data protection issues; for this attempt presupposes that a plethora of personal data is collected, stored, and processed (Hacker 2017, p. 664; Karampatzos 2020, pp. 75–76).^[8]

Within this general framework, Ben-Shahar and Porat (2019, pp. 255 et seq.) advocate the idea of personalising mandatory rules. Their approach is based on the belief that we should create mandatory rules that better fit the needs and expectations of specific contracting parties and reduce the unintended costs of mandatory law. As Ben-Shahar and Porat suggest, that is already happening in the law and business practices within firms personalising the withdrawal right^[9] – to what extent, though, it is hard to tell. According to the authors, a fully personalised law (legal rules) implies that every consumer has their own optimal level of protection, and that the seller charges each consumer a different price. Consequently, Ben-Shahar and Porat (2019, 2021) conclude that personalised legal protection could increase the overall contractual surplus and enable more consumers to enter a contract.

Inter alia, Ben-Shahar and Porat (2019) specifically refer to the mandatory right to withdraw from a contract, suggesting that its protective goal would be better served by a personalised mandatory-rule regime than by a uniformly applicable mandatory-rule regime. For “[s]ome consumers need longer periods to re-evaluate the deal, others can do with shorter. A 72-h right to withdraw from a loan contract may be useless to the weakest of consumers, who are often the neediest, and are also the recipients of the most risky and complicated loan deals. They need more time to overcome the moment-of-purchase confusion. And, conversely, a two-week right to withdraw from online sales may be more than necessary for experienced internet shoppers […]. A uniform duration may be set at the correct “average’ length, but it misfires in individual cases” (Ben-Shahar and Porat 2019, p. 257, also p. 281). Under this prism, if the EU consumers had different cooling-off periods (periods of sober reflection) available at different prices, they would conclude more contracts in the market and increase the overall contractual surplus. The authors rightly point out that, in general, “[p]ersonalized protections also affect redistributive goals, because they eliminate cross-subsidies occurring in equal treatment pools,” for “[w]hen the cross-subsidy is regressive, its elimination is desirable” (op. cit., p. 258). In other words, Ben-Shahar and Porat (2019) advance a personalised mandatory-rule regime that links price differentiation to the degree of legal protection granted to each consumer (meaning that price differentiation goes hand in hand with personalized protection). However, this legal regime is affected by the following disadvantages:

In the first place, creating and implementing a fully personalised withdrawal right is easier to said than done. Such an idea is practically associated with very significant transaction costs, which mostly affect the sellers (Karampatzos 2020, pp. 108–109): for big data must be collected and processed, and thereafter, the sellers must supply different goods with personalised legal protection, along with a corresponding price adjusted to the specific profile of each customer.^[10] Ben-Shahar and Porat themselves (2019, p. 281; cf. also 2021, passim, esp. p. 10) acknowledge the problems related to the technical implementation of their idea, including writing a specific computer code, and developing a complex personalisation algorithm that would enable consumers to get an optimal level of legal protection, i.e., an optimally designed withdrawal right. Assuming that the fully personalised rules do not restrict the autonomy of contracting parties, all these difficulties in design and implementation may generate prohibitive high administrative costs that may outweigh the economic benefits a personalised withdrawal right generates.^[11] In any case, and if such differentiation would ever be practically feasible, it appears that these administrative costs would considerably increase the final purchase price and thus decrease the overall contractual surplus and the total number of concluded contracts.

It also goes without saying that the collection and processing of Big Data raises serious privacy and data protection concerns – an issue that has been generally much discussed in the relevant literature (cf., indicatively, Casey and Niblett 2017; Hacker 2017; Karampatzos 2020, pp. 75–76, 109).

Another general disadvantage of personalised mandatory rules is that they do not enable consumers to waive the protection, i.e., to exclude the withdrawal right; in other words, forced protection and an unwanted price increase cannot be avoided (Karampatzos 2020, p. 108). As a consequence, personalised mandatory rules may still generate destructive or socially unfair cross-subsidies. Namely, consumers who are risk takers and thus do not want the protective shield of the withdrawal right would have to subsidize the provision of the right to those who are risk-averse; this phenomenon may significantly decrease the overall contractual surplus and the number of concluded contracts. A possible counterargument here, though, would be that an efficient fully personalised mandatory-rules regime would take into account the interests of absolute risk takers by totally excluding them from the mandatory imposition of a withdrawal right. However, it appears that, strictly speaking, that would not be a (personalised) mandatory rule.

In light of the above mentioned technical and substantial deficiencies, it seems that (at least at the current level of technological development) one cannot use personalized mandatory rules as an adequate tool for redesigning the withdrawal right under the CRD.

2.2 The Mandated-choice Model

Contrary to the personalised mandatory rules, the mandated-choice model implies that consumers should be given the option to choose between two different contracts, that is, one without a withdrawal right and a lower price, and another one with a withdrawal right and a higher price (Karampatzos 2020, pp. 98 et seq., with further references). If the CRD would provide the consumers with this option on an obligatory (for the seller) basis, it could generate considerable economic benefits. More particularly, the mandated-choice model reflects a reasonable interference with private autonomy, obviously milder than a mandatory imposition of the withdrawal right without a possibility to opt out; as the argument goes, it marries the advantages of the freedom to shape the contractual content to the need to cure existing information asymmetries (Karampatzos 2020, pp. 109–110). It also makes the price of the additional legal protection (embodied in a withdrawal right) visible and transparent.^[12] After having been informed of the withdrawal right and the enhanced protection it offers, consumers would weigh the pros and cons of the legal protection and soberly decide whether the withdrawal right is worth the money or not. Simply by (obligatorily) providing that option to all consumers in distance sales, the mandated-choice model may decrease the socially unjust phenomenon of cross-subsidization – which is an often underappreciated flip side of the extended EU legal paternalism – and increase the overall contractual surplus to a great extent.^[13] For the costs for providing the withdrawal right are borne only by those consumers who opt for the additional legal protection and pay the relevant insurance premium; on the other hand, the consumers who want to take the risk of information asymmetry (the so-called “risk takers” or “lovers”) do not subsidize the provision of the right to those who are risk averse. At the same time, the implementation of the mandated-choice model produces significantly lower transaction costs compared to the creation of personalized mandatory rules. Consumers may check one of the two boxes available online (without the need of engaging any complex computer programs and algorithms), one indicating a contract with and the other without a withdrawal right.^[14]

Hence, under the mandated-choice concept the consumer is free to choose between a contract without a withdrawal right and a contract with the additional insurance of the withdrawal right. The consumer will reach their decision on the basis of their own individual predisposition against risk assumption, their knowledge and experience, as well as the amount of the insurance premium they must pay each time to enjoy the greater legal protection (i.e., the withdrawal right; Eidenmüller 2011a, p. 135). If the consumer opts for higher protection in a distance contract, they are expressing their willingness to buy insurance against the risk of a negative assessment of the good they bought, after having taken possession of it. At that point they might realize that they do not need the good, that it does not meet their expectations, or simply that its (subjective) usefulness or utility does not correspond to the price paid (see Eidenmüller 2011a, p. 135). Having all that in mind, the mandated-choice path seems to be a more suitable tool for redefining the withdrawal right under the CRD. It is, therefore, no coincidence that in the last years this path has found considerable resonance in German legal theory (in German: zwingendes Optionsmodell, optionales Widerrufsrecht) and is embraced by renowned legal scholars such as Eidenmüller (2011a, 2011b) and Wagner (2011).^[15]

However, besides those considerable advantages, the mandated-choice scheme may also create, for both transacting parties (i.e., consumers and sellers), incentives for opportunistic behaviour and entail moral hazards (Karampatzos 2020, pp. 102 et seq.). For instance, the consumer could first buy the good with the withdrawal right, test it, and if it meets their requirements or needs, then exercise the withdrawal right, request a refund of the price paid (200€) and immediately repurchase the good at the considerably lower price (150€) without the withdrawal right. In practice, however, “the price difference between the two products will be much smaller, as is, indeed, the transactional practice when the withdrawal right is voluntarily offered by a seller or provider: in all likelihood, the price for the good with withdrawal right will be, for example, 160€, and for the good without withdrawal right 150€” (Karampatzos 2020, p. 102). Apart from that, though, nothing, prima facie, deters the consumer from temporarily acquiring a product by, practically, borrowing it for use in the context of a specific occasion only, and then from exercising the withdrawal right; such opportunistic consumer behaviours cannot be easily eliminated (once again Karampatzos 2020, p. 103, where further reference to possible regulatory means of countering such behaviours, for instance, by imposing return or depreciation costs on the consumer, and so forth; cf. also Luzak 2013).

Moreover, similar to the personalised mandatory rules, a seller may also behave opportunistically while offering the mandatory choice. Due to their superior bargaining power, a seller may intentionally overvalue the withdrawal right’s option and incentivise consumers to conclude contracts without withdrawal rights – who may always be tempted to do so, even if they are not, by nature, risk takers, possibly motivated by over-optimism.^[16] In this way, a seller may shift all the information asymmetry risk to consumers and behave hazardously to maximize his/her profit; that could decrease the overall contractual surplus and the total number of contracts. Therefore, even though the mandated-choice scheme may solve the critical transaction costs and cross-subsidisation problems, it is still not a perfect tool for redefining the withdrawal right (cf. also Luzak 2013, pp. 30–31).^[17] In fact, it seems that there is room for suggesting a significantly improved and more refined version of the mandated-choice path.

2.3 Risk Allocation Rules

After taking a closer look at the pros and cons of personalised mandatory rules and the mandated-choice model, it seems that the main issue in amending the withdrawal right lies with the perspective of (re)allocation of risk. When contracting parties evenly distribute information asymmetry risk, i.e., when every party may take the risk in line with their own subjective preferences and pay the price accordingly, the incentives for opportunistic behaviour and the probability of moral hazard may drastically reduce. Hence, the question that has to be addressed is how to amend the withdrawal right in order to enable efficient risk (re)allocation between contracting parties.

To answer this question, one must first reveal the preferences – or the predisposition – of a contracting party toward information asymmetry risk. After revealing the individual risk preferences, it is necessary to enable efficient exchange between contracting parties, with minimum transaction costs and cross-subsidies. For that purpose, this paper offers the risk allocation rules, overwhelmingly based on the findings in law and economics theory related to the personalised mandatory rules and the mandated-choice model. Our approach is conceived – at least for the time being – more as a thought experiment; relevant empirical analysis may follow up at a subsequent phase.

2.4 Risk Preferences

The risk allocation rules are specific because, according to the discussion hypothesis adopted in the following lines, they imply that the seller of goods in distance sales should be obliged to provide different offers for the same commodity, containing different prices and cooling-off period ratios. In that sense, a seller could be obliged under those rules to offer the same good at a high price and long cooling-off period, a medium price and medium cooling-off period, and a low price and short cooling-off period. To make the possible benefits of this approach evident, we assume that, under those rules, a seller is willing to provide a 35 % decrease in price along with a 20 % decrease in the cooling-off period length and a 50 % decrease in price along the 55 % decrease in cooling-off period length (compared to the initial offer, i.e., high price and long cooling-off period length).^[18] In this way, the risk allocation rules might help reveal the seller’s preferences toward information asymmetry risk. Namely, the information asymmetry risk, or the risk that the goods will not meet the consumers’ reasonable expectations and that the latter will return the goods to the seller, may be captured and presented as a price/cooling-off period ratio (see Figure 1).

Figure 1:

The Seller’s preferences toward information asymmetry risk. Source: Authors.

Figure 1 shows that a seller is indifferent between offering the same good at a relatively high price and a long cooling-off period (A), a medium price and a medium cooling-off period (B – a 35 % decrease in price and a 20 % decrease in the cooling-off period compared to A), and a low price and short cooling-off period (C – a 50 % decrease in price and the 55 % decrease in the cooling-off period compared to A). The different price/cooling-off period combinations between which the seller is indifferent (they generate the same utility) constitute a seller’s indifferent curve (ICs₁).^[19] Moreover, every seller has an indefinite number of indifference curves (ICs₁, ICs₂, ICs₃ … +∞), and the more remote the curve is from the coordinate origin it represents a higher utility level (U [ICs₁] < U[ICs₂] < U[ICs₃], etc.). Consequently, D (on ICs₂) generates greater utility than any other combination – A, B, or C (on ICs1). Based on the same principle, one could construct indifference curves for every seller in distance sales. However, a seller who is a risk taker (risk lover) would offer relatively lower price/cooling-off ratios, and a seller who is risk averse would offer higher ratios due to the different marginal rates of substitution between the cooling-off period and price.^[20] In any event, due to those risk allocation rules obliging sellers to provide different options to consumers (presumably A, B, and C), the contract would be concluded in accordance with the seller’s risk preferences, regardless of the price/cooling-off period combination consumers choose.

Furthermore, once the seller is obliged to provide different offers for the same commodity and thus expresses their risk preferences (under the risk allocation rules), consumers may also enter the contract following their risk preferences. One may assume that consumers also have their indifference curves, i.e., the different price/cooling-off combinations that generate the same level of utility (see Figure 2).

Figure 2:

The Consumer’s preferences toward information asymmetry risk. Source: Authors.

Figure 2 shows that a consumer is indifferent between purchasing the same goods at a high price and long cooling-off period (C), a medium price and a medium cooling-off period (D), and a low price and short cooling-off period (A). The different price/cooling-off period combinations between which the consumer is indifferent (they generate the same level of utility) constitute a consumer’s indifference curve (ICc₂). Same as in the sellers’ case, every consumer has an indefinite number of indifference curves (ICc₁, ICc₂, ICc₃ … +∞), and the more remote the curve is from the coordinate origin it represents a higher utility level (U [ICc₁] < U[ICc₂] < U[ICc₃], etc.). Consequently, combination B (on ICc₃) generates more utility than any of the combinations represented by C, D, and A (on ICc₂). Based on the same principle, one could construct the indifference curves for every consumer in distance sales.^[21] Nevertheless, it is not necessary to fully reveal consumers’ risk preferences. The fact that consumers are rational (know their preferences and aim at maximizing utility) and that sellers have expressed their preferences (under the risk allocation rules) is enough to enable a Pareto improvement, i.e., more efficient allocation of resources through the exchange by concluding a distance sales contract. One could present that graphically by constructing the Edgeworth box based on Figures 1 and 2.

Figure 3 or Edgeworth Box consists of Figure 2 overlapping Figure 1 and then rotating for 180°. In this way, one could simultaneously observe sellers’ and consumers’ risk preferences. In this case, if a seller would offer one price/cooling-off period combination (which is currently the case under the EU law), for instance, combination A, a consumer would be on his second indifference curve (ICc₂ – medium utility level), and a seller would be on the first indifference curve (ICs₁ – low utility level). However, if the risk allocation rules oblige the seller to offer complementary price/cooling-off period ratios, following their risk preferences (B and C), that could create room for a Pareto improvement or more efficient allocation of resources. In this case, a consumer might opt for the price/cooling-off period combination B and thus switch from the second (ICc₂ – medium utility level) to the third indifference curve (ICc₃ – high utility level), while the seller would remain in the same indifference curve (ICs₁ – low utility level). In other words, under the risk allocation rules, a consumer may significantly increase their level of utility without decreasing seller’s utility.

Figure 3:

Price and cooling-off period Pareto improvement. Source: Authors.

2.5 Risk (Re)Allocation

Besides increasing the utility and, thus, the overall contractual surplus, the suggested rules reallocate the risk and create better incentives for the contracting parties. In our hypothetical example, when opting for combination B instead of A, consumers take more risk because they will have a shorter cooling-off period and less time to reduce the information asymmetry and use the withdrawal right. Moreover, they are adequately compensated for that by the reduction in price. In other words, in combination A, the consumer values more additional increase in risk (shortening of the cooling-off period) than the seller values the decrease. In that case, it is more efficient for contracting parties to opt for B because they take the risk in line with their preferences, and the room for moral hazard is shrinking.

When considering opportunistic behaviour, the risk allocation rules have some disadvantages since there is no guarantee that sellers will reveal their risk preferences in the first place (that they will offer the options on the same indifference curve). Namely, sellers may offer one option that shifts the risk to consumers (for instance, a medium price and short cooling-off period) and other options that are less favourable for consumers (for example, an extremely high price and a short cooling-off period, and even higher price and a short cooling-off period). In this way, sellers may behave opportunistically, avoid the risk allocation rules and pass on the risk to consumers. As above-mentioned, the same could happen with the personalised mandatory rules and the mandated-choice model because of sellers’ bargaining power. However, in the case of the risk allocation rules, one may better address this bargaining problem. When obliging sellers to offer different price/cooling-off period options to consumers, a legislator incentivizes sellers to make their risk preferences more transparent than when resorting to the personalised mandatory rules or the mandated-choice model. In this case, sellers have an incentive to reveal their risk preferences because they have to offer all the options in advance (ex-ante), before the conclusion of the contract, to all consumers, and those options are not negotiable. If a seller ex-ante decides to behave opportunistically and offer one option that shifts the risk to consumers and the other options that are even less favourable, (s)he would be significantly less competitive in the market – and eventually (s)he would be driven out of it. Therefore, consumers would go to the other sellers who fully reveal their preferences ex-ante and offer a better price/cooling-off period ratio. In addition, a legislator may reduce incentives for opportunistic behaviour by limiting the price/cooling-off period ratio through the risk allocation rules. For instance, sellers could be obliged to create at least two additional offers based on the initial one – for every new offer, a seller may increase the price up to 30 % (of the previously offered) and prolong the cooling-off period up to 30 % (of the previously offered), excluding offers without the withdrawal right. In this way, one could eliminate extreme allocation of risk between the contracting parties and reduce incentives for moral hazard and opportunistic behaviour.^[22]

Finally, the most significant advantages of the risk allocation rules compared to the personalised mandatory rules and the mandated-choice model are the low transaction costs and the flexibility in their application. On one side, compared to the personalised mandatory rules, the risk allocation rules do not imply complex negotiations and the use of significant resources (algorithms, computer programs, etc.) before and during the conclusion of the contract. On the other side, compared to the mandated-choice scheme, the risk allocation rules better reveal sellers’ preferences toward risk and thus more efficiently reduce incentives for moral hazard and opportunistic behaviour. In other words, the suggested risk allocation rules take the best of both worlds, i.e., they seem to be a golden mean between the personalised mandatory rules and the mandated-choice model. Those rules offer an almost complete personalization of contracts and efficient risk allocation under low transaction costs. Thus, they reduce cross-subsidization and significantly reduce incentives for opportunistic behaviour as well as the moral hazard of contracting parties. In addition, the risk allocation rules are flexible in their application and could be used in all distance sales, regardless of the industry specificities. All the sellers have to do is offer different price/cooling-off period options, following their risk preferences and being aware of the competitive pressure and regulation in the market. That is the first and the most significant precondition for more efficient exchange and a considerable increase in the number of concluded contracts and overall contractual surplus.

Last but not least, risk allocation rules may share with the mandated-choice model the advantage of the statutory standardisation of the withdrawal right (apart from price and period differences of each available option): the content, the exercise and the legal effects of the right are to be statutorily standardised, and this trait secures the main advantages ascribed to mandatory law – that is, the reduction of transaction costs, the safety of transactions and, in particular, the foreseeability of the legal effects deriving from a contractual relationship or the eventual exercise of a right, such as the withdrawal right (see Karampatzos 2020, p. 99; Wagner 2011, pp. 39–40). Standardisation of the right also limits the ability of the seller to mask or frame their preferred option in such a way that may make it more appealing to the consumer.