ADRonline Monthly

Linking Information Technology and Dispute Resolution

Online Arbitration: Binding or Non-Binding?

Thomas Schultz*


“Non-binding arbitration”. To the vast majority of lawyers, this phrase will sound like a contradiction in terms, arbitration being precisely characterized by the binding character of its awards. Perhaps this position should be reconsidered. But then, perhaps, it is the wording that should be reconsidered. What is certain is that this phrasing sweeps along a rich concept for dispute resolution in cyberspace. This article is not so much on arbitration than on disputes in cyberspace. My hope, however, is that a discussion of online arbitration can illuminate the greatest part of dispute resolution in cyberspace.

In cyberspace, dispute resolution faces a series of fundamental teething troubles. First, in cyberspace parties meet that would often not have met in the offline world, because they physically live in far-away countries or on different continents. Outside cyberspace, ordinary consumers do usually not enter international agreements[1]. In cyberspace, they engage in small or medium transactions, which they would usually not have done in the offline world. In these cases, courts are too expensive—mainly because of costs of filing, travel and legal counsel—and they are too slow. Courts are often an economically unreasonable medium to solve disputes arising out of cyberspace. The parties to such small or medium-sized disputes in cyberspace will often have strong economic incentives not to go to court, leaving the dishonest party with a victory. Courts, for such parties with disputes of this kind, generate plainly prohibitive costs.

Second, alternative dispute resolution methods except arbitration are often not effective enough, because they produce case outcomes that are not legally binding enough. Such methods only engender contracts—if they work. To be enforceable by state authorities, contracts again require a judgement—which is too expensive. As it is expensive, the winning party will have only limited incentives to go to court, and the losing party will also have only limited incentive to obey the contract, because it is unlikely to become enforceable.

Third, arbitration could provide this effectiveness, because the recourse to courts is only minimal and therefore much less expensive. But from this legal effectiveness follow other difficulties. Arbitration requires the parties to give up rights, which does not induce trust and which is the reason why arbitration currently still faces a series of legal obstacles. For consumer claims, where the cost problem is greatest, there is for instance a problem of arbitrability under a number of laws. Although it is probably correct to say that most of these legal obstacles are merely flaws in the legal system (the requirement of a writing on paper does not protect the parties; due process can certainly be provided in a fast-track procedure; consumers are not necessarily better protected in court than in arbitration; requiring an original award does not make sense in an electronic environment), these obstacles are still there.

My aim in this paper is to convince you that electronic commerce is best left to self-regulation and to dispute resolution systems that are both effective and not faced with the legal obstacles posed by the national legal systems. My main argument will be to show that, in cyberspace, for small and medium-sized cases of electronic commerce, non-binding arbitration may in fact be more binding than traditional arbitration, because it is effective without being subject to so many legal obstacles.

In other words, I advocate self-regulated dispute resolution, although I must immediately make clear that such dispute resolution systems are only best as effectiveness is concerned, not as legitimacy is concerned—which is another important problem that cannot be addressed in the balance of this paper.

I will argue that such a self-regulated dispute resolution system is online non-binding arbitration. I will sketch the legal problems that online binding arbitration faces (III), and then introduce the ways in which non-binding online arbitration can be made binding de facto (IV). Before that, I must make a brief survey of the activity of online arbitration (I), and categorize the forms of online arbitration (II).

I. Online arbitration: what is out there?

Monitoring an activity in cyberspace is often not easy, because there is too much information on the Internet and because what you see is not always what there really is. The assessment in this paper of what is out there is therefore based on a variety of sources: the existing reports on the state of the art of ODR[2], updated by a series of web visits, colloquia and interviews. Still, I must admit, the following information may not be absolutely accurate, but they probably provide a picture realistic enough to ground the inductive reasoning leading to the concepts that are developed later.

On the Internet, one can find over twenty-five providers of online dispute resolution offering a service either labeled ‘arbitration’ or resembling the arbitration procedures that we know from the offline experience[3]. The providers usually declare that they offer binding as well as non-binding arbitration, though some of them restrict their services to the non-binding form[4].

There seems to be a fairly strong correlation between the binding character of the outcome and the caseloads. Non-binding awards are rather frequent, while their binding counterparts are quite infrequent (or not reported to the full extent of the phenomenon). For instance, under ICANN’s Uniform Dispute Resolution Policy for domain names (UDRP)—which I will consider the epitome of non-binding online arbitration—over six thousand decisions have been rendered[5], whereas the Chartered Institute of Arbitrators—which probably holds the lead in the field of online binding arbitration—has delivered approximately 70 awards[6]. Comprehensive statistics on online arbitration, however, are not available yet, which is probably due to the fact that the parties to arbitration and the institutions providing it consider such procedures private and confidential.

The other important part of the picture of online arbitration is its material setup, which strongly contributes to its success. The parties, in such procedures, usually communicate by emails, by web-based communication tools and—not yet but certainly  soon—by videoconferences. The fees for online arbitration are often charged on an hourly basis, and then range from 50 to 250 USD per party and per hour. But under the UDRP, the fees range from 1500 to 4000 USD—depending on the number of domain names at stake and the number of panelists. They are borne by the complainant, except when the respondent chooses a three-member panel. In the online arbitration schemes of the Chartered Institute, the fees range from approximately 15 USD for a dispute worth less than 300 USD—in the Which?WebTrader scheme[7]—to 250 USD for a dispute worth between 7’500 USD and 15'000 USD—in the ABTA scheme[8]. There are usually no time limits in arbitration, but when there are, they vary between 4 hours and 60 days. Under the UDRP, there is a series of time limits, which bring the procedure to an average of 2 months. At the Chartered Institute, awards are delivered after an average of 7 weeks[9].

After an empirical survey, a first observation comes to mind. Out there, in cyberspace, time passes quickly. Providers of dispute resolution services burgeon, develop, search for the right business model, usually do not find it, and often soon cease activity. However, out of this bustling activity a particular form of solving dispute is emerging: online non-binding arbitration. These are the data that can be found by empirical research. I must now try to conceptualize facts.

II. The forms of online arbitration

Arbitration is a process where a third party chosen by the parties, or nominated by the institution chosen by the parties, renders a decision on a case while applying fundamental procedural principles. Traditionally—in the offline world—arbitration resolves a dispute by making an enforceable decision[10]. This is the binding form of arbitration, a process whose decisions, by operation of law, are enforceable by state authorities[11]. The ideology of binding arbitration is that it constitutes a form of private judging, a substitute for court litigation[12]. Let us consider this as the ideal type of arbitration[13].

In addition to this “perfect” form of arbitration, there is a non-binding form. ‘Non-binding’, in this context, means that the decision is not enforceable by state authorities on its own. At the time the decision is rendered, something is missing to make it enforceable by operation of law. Such a decision may become binding afterwards, when made part of a consent award or a settlement agreement. The ideology of non-binding arbitration is that it constitutes negotiation assistance[14], a form of mock trial. It this respect, it resembles the last phase of some strategies of mediation, where the mediator delivers an opinion, making a decision—which has no binding effect—in lieu of an agreement[15]. As it gives both sides a balanced view of the case and a glimpse at the possible outcome in court of in traditional arbitration, it is hoped to carry great weight in subsequent settlement discussions. If the parties acquiesce to it, the decision becomes part of a settlement agreement or a consent award and can then be enforced[16]. It may also be unilaterally binding, in the sense that only the stronger party—usually the seller—is bound by the decision, and only the other party may choose to acquiesce or not[17].

The legal nature of non-binding arbitration depends in principle on the applicable lex arbitri, which in turn depends on the seat of arbitration. In the United States, where non-binding arbitration is not infrequent even offline[18], such process is likely to be characterized as arbitration[19]. In Europe, however, it may be inconsistent with the narrower conception of arbitration, as one generally considers that a decision constitutes an award, and the process leading to it an arbitration, if the parties intend it to be binding like a court judgment[20]. All the same, as it may produce a consent award[21], it could be characterized as arbitration. This issue is anyway not as important for this exposé as the next type of arbitration.

In addition to this first form of non-binding arbitration, there is another, a second form of non-binding online arbitration. The difference between the two forms of non-binding arbitration is that the form that has just been described may lead, in the end, to a binding award, whereas the form that will now be described never produces a binding award. To be more precise, the first form of non-binding arbitration is optionally binding, whereas the second is truly non-binding; but both are non-binding at the time the decision is rendered (see Figure 1).

The term ‘binding’ being here understood as the capacity to be enforced by public authorities in the same manner as a judgement can be enforced. This is important because even truly non-binding arbitration is not binding merely from the point of view of state authorities, but not from the perspective of the social sub-system they act within[22]. From the point of view of the social sub-system—for instance a given marketplace, or the domain name system—, truly non-binding arbitration decisions are as binding—as coercive—as a judgment or a normally binding arbitration award. The truly non-binding decisions are coercive—or binding—because they can lead to self-enforcement[23], that is to enforcement by private authorities without the intervention of public authorities[24].

The ideology of this form of arbitration is more than negotiation assistance, it switches back to a role of private judging, a form of private judging very specific to a given social sub-system and independent from courts. The most notorious example of that form of non-binding arbitration is the UDRP[25], which creates a sub-system around the concerned domain names, with their own transnational legal regime.

Figure 1. The ideal-type of arbitration is binding. There are less ‘proper’ forms of arbitration which are not binding.

But why would one use something else than binding arbitration? First, there is a meme that marks many ODR skeptics—an idea that defines the role of ODR in general. The goal of ODR is not so much to settle a dispute, but rather to enhance consumer confidence in electronic commerce[26]. From this point of view, online arbitration is a form of external client service: the customers should see a system related to their dispute, which looks good and reassuring, and which would not put at stake—should it be used—too much time and money. If such a system is binding and the customers feel they are giving up rights, it will trigger doubts and miss its goal. If a dispute occurs, so the thinking goes, either both parties are satisfied with the outcome and abide by it, or only very little time and money is lost.

Second, there are many legal barriers to binding online arbitration. There is the arbitration clause and the form of the award. And there is the binding character of the outcome, which can limit the business contexts of the process, as the scope of arbitrability is restricted under some arbitration laws to protect the weaker party, which may be a consumer in an electronic commerce transaction.

On the other hand, it may be considered that such a non-final decision does not put an end to the procedure, that it just adds another layer of procedure before the parties are really able to vindicate their rights[27]. In other words, that non-binding arbitration is not effective enough.

There may be a third way. Non-binding arbitration could be used, but with additional features that make it more effective. The parties would not give up their rights—or maybe would give up fewer rights—while participating in a process that provides some effectiveness. And most legal obstacles could be removed. This is what I will explore in the balance of this paper.

III. Legal obstacles to binding online arbitration

In the current state of legislation and practice, resolving a dispute through online binding arbitration—arbitration proper—faces a number of legal obstacles. Such obstacles arise throughout the process: first the agreement to arbitrate, then the arbitration procedure, and finally the recognition and enforcement of the award.

Agreements to arbitrate online face problems concerning their validity, their evidence and their enforceability. Agreements to arbitrate online are usually entered into online. Unfortunately, most national laws and international convention still require the arbitration agreement to be in writing[28]. Only the UNCITRAL Model Law (Art. 7 (2)), the German Arbitration Act (§ 1031 (1)), the English Arbitration Act (Sect. 5 (6)), and the Swiss PIL Act (Art. 178 (1)) consider that the agreement is in writing when recorded by electronic means[29]. Most laws, however, do not. The current wording and interpretation of the New York Convention (Art. II (2)), for instance, do not include the agreement being recorded by electronic means[30]. In addition, the arbitration agreement, which is usually part of the general terms, must be incorporated into the contract in a certain manner. It requires—I am simplifying brutally in my consideration of that issue here—an express wording stating that the contract is governed by the general terms, which a user exercising normal care must be able to review[31]. This issue of incorporation does not have to be further discussed here, as it is only a slight modality of the conclusion of the contract and does in no way pose any serious obstacles on online binding arbitration. My aim, however, is simply to outline some sensitive elements in the online conclusion of an arbitration agreement, though one must certainly recognize that anything written can be recorded either on paper or on electronic storage—in other words, that an electronic document is in most circumstances the functional equivalent of a paper document. This, however, leads to some technical difficulties as its character as a probative evidential record is concerned.

To be a valid arbitration agreement, the electronic document must include the identity of the parties, the agreement itself (i.e. the offer and the acceptance), and the content of the agreement (i.e. the specific terms and the general conditions). This information must be stored in a manner that allows its accessibility for further reference and its admissibility as evidence. In other words, this information must be stored using a technology which permits long-lasting compatibility[32] and which excludes any serious risk of manipulation of the stored data[33]. There seem to be ways to render electronic documents as difficult to forge as paper documents. They are for instance invisible watermarks on electronic documents[34], the use of a cybernotary[35], and, in some circumstances, digital signatures, which authenticate the author and the content of a message[36]. Routine operating procedures—such as backup audit files—used by some larger companies apparently also provide reliable records of email communications[37]. Be it as it may, evidence raises technical issues, which cause legal controversies as the technological means employed may be characterized differently under different laws.

If the case involves a consumer, the arbitration process is in addition faced with the question of the arbitrability of the dispute[38]. Agreements to arbitrate, even when formally valid, may not be enforceable—courts might not refer the parties to arbitration. Many arbitration laws consider that situations where the parties have substantially different bargaining powers are non-arbitrable subject matters—they go against public policy. Such laws seek for instance to protect tenants, employees, or—more relevant to us—consumers[39]. In the United States, consumer arbitration clauses are held enforceable unless they are procedurally unconscionable (which requires a finding of unfair surprise or lack of notice) or substantially unconscionable (they lead to an unreasonably high expense of arbitration fees imposing a burden on the consumer)[40]. European laws are generally less favorable to consumer arbitration and pose higher burdens[41]. The EU Directive on unfair terms in consumer contracts, for instance, declares unfair those clauses that are “excluding or hindering the consumer’s right to take legal action or exercise any other legal remedy, particularly by requiring the consumer to take disputes exclusively to arbitration not covered by legal provisions”[42]. “Not covered by legal provisions”, however, has been interpreted quite differently in the various European countries and this opposition towards arbitration must not be overstated. The EU might even be moving towards a system of accreditation that may set the conditions under which the consumer could exclude its right to litigate[43]. Whatever the case may be, the fact is that one can currently assume that pre-dispute arbitration clauses excluding the professional and the consumer’s right to take the disputes to court—thus bilaterally binding—are more likely to be held unenforceable than not in most European countries[44]. As the near future of ODR is likely to involve primarily B2C transactions[45], these current arbitration laws are obstacles to the development of online binding arbitration. This problem may be overcome by using unilaterally binding arbitration agreements, also called optional arbitration clauses—clauses that bind only the professional, leaving the consumer free to decide whether to take the dispute to arbitration or court. Then again, in some cases, unilaterally binding clauses have also been held not enforceable in the US, on the grounds of the doctrines of mutuality and consideration—such clauses are not mutually binding, and if one of the parties is not bound, none is[46].

An online binding arbitration procedure also faces obstacles related to its communication architecture and requirements of due process[47]. As in any other commercial activity on the Net, the parties demand an architecture of security in online arbitration. Such architecture would typically have to provide (1) authentication, to ensure the identity of the parties at the time of the transaction and during the ODR procedure; (2) authorization, to ensure that the other party and the arbitrator are sanctioned for their function in the process; (3) privacy, to preserve the confidentiality of the parties at least during the proceedings; (4) integrity, to ensure that the contents of a record have not been altered; and (5) non-repudiation, to show that a file or program has been entirely transmitted to the buyer in case of an online performance of a contract[48]. In order not to violate due process requirements, the proceedings must in addition make use of a communication architecture that allows proportionate and effective means for the receipt of evidence. The problem is that one of the fundamental characteristics and advantages of online arbitration—that is, of online dispute resolution in general—is speed and low costs. Speed implies simplified procedures and less formalism and low costs make a fully-fledged architecture of security economically difficult to sustain. These demands on arbitration—due process and security on the one hand, speed and low costs on the other—work against each other. The trouble, one must keep in mind, is that too little opportunities to be heard are a possible jeopardy to due process[49]. But exactly how much one can expedite an arbitration before violating due process requirements—the award then running a risk to be set aside in court—is nowhere clearly defined[50].

Finally, online awards face problems of effectiveness; that is, they may not be recognized or enforced if they are rendered online. Here, online binding arbitration faces technical as well as legal problems. Computer scientists, and indirectly lawyers, are faced with the delicate questions of who should store and send the award to the enforcement authority. Consider the trouble that there may be a problem of falsification of the award if it is sent by the (allegedly) winning party, as the document has been in its electronic storage. Even if it were technically possible to ensure its authenticity by ‘freezing’ the document in its electronic repository, online arbitration institutions will have to use much persuasion to inspire trust in such a solution. The award could also be sent by the arbitral institution or the arbitrator(s), but they may no longer be available at the time of the recognition or enforcement. Another solution is to have the award sent by a trusted third party, such as a cybernotary or a centralized registry record[51]. This last solution may be the best, because such a third party could easily the state-owned, and would thereby run less risk of ceasing activity.

In addition to this issue, consider how the award may be notified to the parties. The trouble is that current email protocols are not able to produce proofs of receipt, and counting on the losing party to collaborate in her notification is probably too optimistic.

Lawyers are also faced with the problem that, under the New York Convention (Art. IV (1)) traditionally interpreted, the party moving for enforcement must provide an award that is in writing[52], signed by a majority of the arbitrators, and that is either the authenticated original or a duly certified copy thereof[53]. Of course, technically theses conditions could be met if electronic documents are recognized to qualify as writing and if electronic signature is used—because it can authenticate the sender and the content of the message—, but these solutions do not correspond to the current wording of the New York Convention, nor to its common interpretation[54]. An obvious solution—and currently certainly the safest—would be to use a printed version of the arbitral award[55]. It could easily be signed by the arbitrators who could check the content of the award—thereby excluding risks of electronic falsification—, it could be properly notified to the parties by snail-mail, and the signed print-out would clearly constitute the original award. This solution, however, requires more time, more costs, and a physical storage of the signed award—which may be an undesirable burden in small cases[56].

IV. Rendering non-binding arbitration effective

In truly non-binding arbitration, these problems do not exist, as this form of arbitration seeks neither the recognition of its agreements, nor the compatibility of its procedures with requirements of due process, nor the recognition of enforcement of its decisions by state authorities. From its non-binding character, however, also follows an obvious problem: the enforcement of the decision. If state authorities do not back up these decisions with their enforcement power, how can they be effective? Where is the incentive for the losing party to perform?

The answer is that the winning party may hope for unforced compliance or the dispute resolution system may use a mechanism that produces a binding force alternative to the enforcement power of the state.

Voluntary compliance to a non-binding decision rendered after fair and trust-inducing proceedings, no matter whether online or offline, is actually not unlikely. Non-binding arbitration can play the role of both an “advisory opinion” and a place to vent. As an advisory opinion, it helps the parties to reassess their own opinion on their position; they can test their arguments in a “trial run” and evaluate the likely outcome of adjudication[57]. As a place to vent, it may provide some catharsis; it may help alleviate anguish and aggression through expression and revelation. For both of these effects to take place, it is important that the parties feel that they have obtained a fair hearing—which supposes the use of the appropriate communication tools[58]—and that they have been handed down a decision from an expert third party who is truly impartial. The most striking examples are UDRP decisions, whose compliance rate is extremely high, although the UDRP is a somewhat special case, as will be discussed below.

The alternative binding force of a dispute resolution system can rely on either particular incentives to perform or on a mechanism that allows an enforcement of the decision by private authorities—so called self-enforcement. Both of these structures rely on a common lever: the control of a resource that is valuable to the parties. In the offline world, law enforcement relies, in last resort, on the physical control of the freedom of the subjects of the law by the police, one of the organs of the state. In cyberspace, enforcement of non-binding arbitration decisions relies on the control of the money of the parties, their reputation, or their domain name, by a private authority linked to the ODR provider[59].

One way to create incentives to perform is to trustmark web traders[60]. A trustmark is a logo displayed on the website of the trader, which informs the customer that the trader has committed to complying with a number of qualitative standards or best business practices, including for instance redress mechanism such as non-binding online arbitration[61]. The trader can for instance commit to comply with all decisions of an online non-binding arbitration procedure[62]. If the trader does not abide by and perform the decision—if he does not comply with this best business practice—the trustmark is removed. This of course presupposes that the removal of the trustmark is controlled by the ODR provider, either directly or indirectly, by networking with the controlling entity[63]. The losing party’s incentive to perform the decision follows in this case from the risk of losing the trustmark.

How strong this incentive would actually be is difficult to assess, because the importance of a trustmark for a website is not easy to evaluate. A trustmark is meant to increase the trust and confidence of customers in a web trader[64] and most people using the net declare that they would be reassured by a trustmark[65]. Many authors emphasize their importance for e-commerce[66] and even several governments encourage them[67]. Nevertheless, how important this really is for the trader, whether the trader will deem it preferable to lose whatever the decision makes him lose or lose the trustmark, remains doubtful.

Another way to create incentives to perform is to threaten to exclude participants from a marketplace, by using technological tools—for instance a website secured by passwords—to deny them access. In a large marketplace such as eBay, accessing and participating in the business of the place can be economically important for the parties. It may be important to the point that being evicted from the business of the marketplace would carry heavier economic consequences than abiding by an unfavorable decision of an ODR provider. Under a threat of exclusion, the most (economically) reasonable decision for a party may simply be to comply with the outcome of the dispute resolution procedure and keep on doing business[68]. This solution, however, would hardly work with one-shot players.


Mechanisms that allow to self-enforce decisions are for instance escrow services; judgment funds; transaction insurance mechanisms; privileged links with credit card issuers; and particular technological tools.

Escrow services operate by holding in escrow the buyer’s payment until the buyer has received and inspected the merchandise[69]. The buyer sends her payment to the escrow account, not to the seller. Once the funds are verified and secured, the escrow service notifies the seller to ship the item to the buyer. The escrow service then tracks the shipment and, after an inspection period (typically two days), releases the payment to the seller, unless the buyer notifies the service that there is a problem. If the buyer is not satisfied, the escrow service releases the payment back to the her. The escrow company acts as a secure payment intermediary, which holds an account on which the money transits. Both parties must register with the escrow service and enter into an escrow agreement before the transaction. This agreement often specify that a dispute between the buyer and the seller will be resolved by arbitration[70].

In the system of judgment funds, a fund is constituted beforehand by contributions of the prospective parties—in the case of B2C dispute resolution by the affiliated web traders. When a decision is rendered, the awarded sum of money is taken from the judgment fund. If the ODR provider controls this fund, he can enforce the decision on his own. Such systems are used outside electronic commerce, for instance by the United Nations Compensation Commission[71] and the German Forced Labor Compensation Program, which operates partly online[72].

A transaction insurance mechanism is a form of money-back guarantee, which intervenes when then dispute could not be solved by an agreement between the parties. The difference with a judgment fund is that it does not rely on prior contributions of the parties. Think for instance of the following scheme: a dispute occurs; the parties negotiate but do not reach an agreement; a hypothetical mediator tries to help but does not achieve better results; an arbitrator is selected and renders a decision; if the winner is the buyer, he gets his money back, the amount being provided directly by the ODR provider or a related entity; this entity of the ODR provider reclaims this sum of money from the losing party[73].

A privileged link with a credit card issuer could also be an instrument providing for self-enforcement. Under the laws of most countries, the cardholder has no liability for non-authorized uses of his credit card—unless he has acted negligently[74]. However, in case of an authorized use of his card, the cardholder will be charged, no matter whether he receives the goods or not. The cardholder will simply be charged when he authorized the payment. In the US and the UK, the situation is different[75], as there exists a mandatory chargeback regime in cases of certain problems in the underlying transaction. Under US and UK law, the cardholder has a right to be recredited with the sums paid if the goods or services have not been delivered as agreed[76]. In other countries, such a right to charge back in cases of non-performance or bad performance could be contractually provided for by the card issuer and the merchant. And this right could be submitted to the decision of a provider of online arbitration. Imagine a provider of online arbitration making a contract with a credit card issuer, who in turn makes a contract with a merchant. The terms of the contracts would stipulate that the credit card issuer will charge back the merchant if the arbitrator has decided so. Such a contract may be accepted by the merchant because (1) he would become more attractive to online consumers, and (2) card issuers—who are actually payment intermediaries—have a significant leverage on merchants: they are the payment providers. If the merchant wants to get his money—which both honest and fraudulent merchants want—, he depends on the card issuer. The resource that is valuable to the merchant—the money—is controlled by the payment intermediary. The intermediary could even increase the merchant’s incentive to obey by threatening to terminate its services to specific merchants[77].

Technological tools allow, as I said, to exclude participants from a marketplace. But in some specific cases, particular technological tools can also be used to enforce the decision directly, not just to threaten the parties. Such a mechanism exists under the UDRP. Ten days after the decision is rendered—subject to a party bringing court proceedings—, the domain name is canceled or transferred[78] by the registrar of the domain name, who is contractually bound to do so. The technological tool which renders this self-executing mechanism possible is the control by ICANN of the database which converts domain names into IP addresses: if a domain name registrar wants his domain names to be converted into IP addresses, he has to accept the conditions set by ICANN[79], among which to commit to enforce all decisions rendered by an ICANN-approved dispute resolution institution[80]. If the dispute is of low economic value, it is unlikely that the losing party would seek to litigate after a decision has been self-enforced[81].

This mechanisms can obviously also be combined. One could imagine, for instance, a marketplace using or offering (1) a reputation system based on trustmarks attributed to the businesses that comply with the marketplace’s code of conduct; (2) a judgment fund for its large affiliated business ventures; (3) an escrow service to secure the payments of the consumers; (4) a transaction insurance scheme to offer a backup for its most suspicious parties; (5) a privileged link with credit card issuers for those parties who simply do no longer wish to deal with the marketplace; and (6) its power of exclusion in cases of repeat-players.

V. Conclusion: Non-binding arbitration, an instrument of legal pluralism

Before really concluding, I must quote two authors who made a point that comes parallel to the effectiveness of non-binding arbitration.

In an article on the importance of studying how law and cyberspace connect—this connection shedding light on other regions of the law—Lawrence Lessig argues that computer code resembles contracts, but code does not equate to contracts[82]. They differ in that law by contract (as opposed to code) requires the intervention of a court to become enforceable; it has a number of critical components to be enforceable through court[83]. The constraints of contracts rely on exogenous factors—a reason why legal realists have taught that contract law is actually public law—and these factors are controlled by the (public) enforcers when they make a judgment about whether this obligation should be enforced. Code does not. Code carries self-imposed constraints—the constraints of code rely on endogenous factors. Code-makers can give an order, an order backed by a threat imposed by code. Lessig then somewhere else concludes that persons affected by code are compelled to comply with the constraints of code more completely than traditional enforcers would compel them to comply with conventional legal rules[84]. In other words, code generates a self-enforcing regime.

In a profound article on the regulatory schemes of cyberspace, Henry Perritt defines the de facto control over a valuable resource by private persons or entities as a source of private regulation[85]. The idea is that if a private entity has the authority to exclude others from a resource—for instance a domain name[86]—it has de facto the authority to regulate the access and use of the resource. If the resource is more valuable than the regulation is troublesome, people or entities will accept to submit to the regulation—the control of the resource entails the de facto power to regulate. The control of the resource also carries self-imposed, endogenous constraints. Resource controllers can give an order, an order backed by a threat to deny access to the resource. In other words, resource control generates a self-enforcing regime.

In sum, the ability to produce law may very well be a consequence of brute force. It then follows that the entity that can do without state authorities is able to produce its own law The examples show that neither code-makers nor resource controllers need state authorities to make their decisions enforceable. Such a situation allows self-regulation and thereby contributes to legal pluralism, to the plurality of legal orders.

In the case of non-binding arbitration and electronic commerce, one could think of a legal order based on a marketplace—such as eBay. First raise borders around this marketplace, by passwords and other tools controlling access. Then make decisions of arbitrators effective inside this legal order. For this purpose, non-binding arbitration would have the advantage of its independence vis-à-vis state authorities. Think of the legal obstacles sketched out above—which are difficult to secure, I have tried to show—as critical components of enforceability, as exogenous factors of constraint, which are controlled by the public enforcers. Non-binding arbitration does not rely on such exogenous factors, it can do without states and courts—when it is equipped with self-enforcement systems (trustmarks or other reputation-based systems; threats of exclusion; escrow services; judgment funds; privileged links with credit card issuers; particular technological tools; a transaction insurance scheme). The marketplace with non-binding arbitration enforces the decisions itself. It carries out a self-imposed constraint. It operates free of the limitations of binding arbitration. The decisions do not have to be enforceable through courts; they must simply be enforceable in the legal order of the marketplace.

Finally, as legal pluralism is concerned, compare Hart when he describes law as an order backed by a threat directed at primary behavior[87] and the normativities observed on eBay by Ethan Katsh, Janet Rifkin and Alan Gaitenby: “[a] marketplace could … rely on an arbitration process … and use the threat of exclusion as the mechanism for enforcing the terms of the ruling. … [W]e increasingly felt that eBay could be considered to be a jurisdiction in itself, a legal authority in itself, an entity that might even be considered to be able to exercise a loosely defined sovereign power…”[88]. This is exactly how online non-binding arbitration could work: independent from state and courts, with its own system of enforcement, and its own threats to back its orders. Maybe this mechanism could illustrate how dispute resolution in cyberspace should best be implemented.


*, Geneva E-Com / E-Law Research Project on Online Dispute Resolution (University of Geneva), <>. The errors and omissions in this paper remain the sole responsibility of the author.

[1]     Lawrence Lessig, Code and Other Laws of Cyberspace, New York, Basic Books, 1999, 197 (“[I]nternational agreements for the most part are agreements between sophisticated actors. Before cyberspace, ordinary consumers were not international actors”).

[2]     ICC, “Business-to-Consumer and Consumer-to-Consumer Alternative Dispute Resolution (ADR) Inventory Project. Summary Report”, released on May 14, 2002; Consumers International, “Disputes in Cyberspace. Update of online dispute resolution for consumers in cross-border disputes”, November 2001, and Consumers International, “Online dispute resolution for consumers in cross-border disputes – an international survey”, by P. Lawson, December 2000; Thomas Schultz, Gabrielle Kaufmann-Kohler, Dirk Langer, and Vincent Bonnet, “Online Dispute Resolution: The State of the Art and the Issues”, E-Com Research Project of the University of Geneva, Geneva, 2001,, (updated regularly since).

[3]     The following entities were found to be offering online arbitration:        the American Arbitration Association, the Association of British Travel Agents (ABTA); BBBOnline; the Chartered Institute of Arbitrators (providing ODR for the Ford Motor Company, the Association of British Travel Agencies (ABTA), Which?WebTrader, the UK Musicians’ Union, and soon for the EEJ-Net); the CiberTribunal Peruano; the Commercial Initiative for Dispute Resolution; Cyberarbitration; Cybercourt; eResolution; the Honk Kong International Arbitration Center; IntelliCOURT; iCourthouse; Internet Ombudsman; MARS; NovaForum; ODR.NL; Online Resolution; the Resolution Forum; SettleTheCase; SmartSettle; SquareTrade; Trusted Shops; the Virtual Magistrate; Web Assured; Web Dispute Resolutions;; WebMediate; Word&Bond and the four ICANN-approved providers (WIPO; the National Arbitration Forum; the Asian Domain Name Dispute Resolution Center; and the CPR Institute for Dispute Resolution). This list is probably not exhaustive.

[4]     Exclusively non-binding arbitration is for instance provided by the Asian Domain Name Dispute Resolution Center; the CPR Institute for Dispute Resolution; iCourthouse; the Virtual Magistrate; and WIPO.

[5]     In June 2002, statistics showed that 3.700 cases had bee submitted to WIPO,, over 2.400 to the National Arbitration Forum, > search without entry, and some 50 the CPR Institute, The Asian Domain Name Dispute Resolution Center (approved on February 28 2002) does not publish statistics yet. eResolution, which ceased activity on November 30, 2001, had administered approximately 300 disputes.

[6]     This figure has been provided by Gregory Hunt, Manager of the Dispute Resolution Services of the Chartered Institute of Arbitrators, during an interview in London, on September 19th, 2002.

[7]     See the fees table of the ODR program for Which?WebTrader at

[8]     See the fees table of the ODR program of the Chartered Institute for the Association of British Travel Agencies, at

[9]     See the Guidance Notes for both the ABTA and the Which?WebTrader schemes, mentioned above.

[10]    Alan Redfern and Martin Hunter, Law and Practice of International Commercial Arbitration, 3rd ed., London, Sweet & Maxwell, 1999, 3-4; Charles Jarrosson, La notion d’arbitrage, Paris, L.G.D.J, 1987, 110; Charles Jarrosson, “Les frontières de l’arbitrage”, Rev. Arb., 1, 2001, 5-41; Gabrielle Kaufmann-Kohler and Henry Peter, “Formula One Racing and Arbitration: The FIA Tailor-Made System for Fast Track Dispute Resolution”, 17 Arb. Int’l. 173 (2001) (on the fundamental procedural principles); Andreas Bucher and Pierre-Yves Tschanz, International Commercial Arbitration in Switzerland, Basle and Frankfurt on the Main, Helbing & Lichtenhahn, 1989, 26.

[11]    ‘Binding’, here, refers to the awards and means that it is enforceable by state authorities. It could also refer to the arbitration agreement, which could be either binding or optional, but this is not my concern here.

[12]    Albert Jan van den Berg, The New York Arbitration Convention of 1958: towards a uniform judicial interpretation, Deventer, Boston, Kluwer Law and Taxation, 1981, 44; Philippe Fouchard, L’arbitrage commercial international, Paris, Dalloz, 1965, 1, 30, 31, William S. Holdsworth, A History of English Law, London, Sweet & Maxwell, 1964, Vol. XIV, 187, and Michael J. Mustill, “Arbitration: History and Background”, Journal of International Arbitration, 6, 1989, 43.

[13]    Such a statement may surprise European lawyers, because one usually considers that a process qualifies as arbitration if it meets a series of conditions (there must be an agreement to arbitrate, the arbitrators must be chosen, directly or indirectly, by the parties, the arbitral tribunal must apply fundamental procedural rules and render a decision ending the dispute, and this decision must be binding and enforceable). This position, however, is in essence a positivist approach. Following the current paradigm shift in legal theory, I will try and consider that a process is arbitration if it sufficiently resembles the defined ideal-type.

[14]    “Should Mediators Evaluate?: A Debate Between Lela P. Love and James B. Boskey”, 1 Cardozo Online J. Confl. Resol 1 (1999-2000); and Alan Scott Rau, “Contracting Out of the Arbitration Act”, 8 Am. Rev. Int’l Arb 225 (1997), 242 (comparing non-binding arbitration with “other formal “reality testing” devices”).

[15]    Cf. on the recommendations under the rules of ECODIR, Alexandre Cruquenaire and Fabrice Patoul, “Le développement des modes alternatifs de règlement des litiges de consommation : Quelques réflexions inspirées par l’expérience ECODIR”, 7 Lex Electronica (2002),, § 75-80 (At the end of a mediation procedure, the mediator renders a recommendation. The parties can accept the recommendation either before or after it is rendered. In both cases it has the binding character of a contract).

[16]    See for instance, concerning a patent dispute, Scott H. Blackmand and Rebecca M. McNeill, “Alternative Dispute Resolution in Commercial Intellectual Property Disputes”, 47 Am. U.L. Rev. 1709 (Aug. 1998), 1714.

[17]    The FordJourney and Web Trader schemes, for instance, produce awards that are binding on the business, but not on the consumer. See FordJourney at (“It should be noted that the Service is legally binding only on FordJourney, and the Claimant can reject the Award and proceed through the courts if they are dissatisfied with the outcome”) and Web Trader at (similar wording).

[18]    The Judicial Improvements and Access to Justice Act (Pub. L. No. 100-702 (H.R.2127), 28 U.S.C. §§ 651-658 (1994)), for instance, authorizes 10 districts to practice non-binding arbitration as a judiciary dispute resolution program.

[19]    Rau, “Contracting Out of the Arbitration Act”, 243 (“[s]uch agreements for non-binding arbitration have been held to be within the Federal Arbitration Act for the purposes of stays or orders to compel under § 3 and 4”), cf. 240-241 (advocating its validity but reporting that “a number of state courts have indeed held such clauses in insurance contracts to be unenforceable. For the Supreme Court of Minnesota, for example, the trial de novo provision “would result in complete frustration of the very essence of the public policy favoring arbitration”.

[20]    For Italian law, Piero Bernardini, Il diritto dell’arbitrato, Rome, Laterza, 1998, 17. For French law, Jarrosson, La notion d’arbitrage, 162 and Alan Scott Rau and Catherine Pédamon, “La contractualisation de l’arbitrage: le modèle américain”, 2001 Rev. Arb. 453, 461, 482. For German law, Klaus Peter Berger, Internationale Wirtschaftsschiedsgerichtsbarkeit, Verfahrens- und materiellrechtliche Grundprobleme im Spiegel moderner Schiedsgesetze und Schiedspraxis, Berlin/New York, De Gruyter, 1992, 53-54. For Swiss law, Felix R. Ehrat, Commentary ad Article 176 PIL, Kommentar zum schweizerischen Privatrecht, Internationales Privatrecht, Heinrich Honsell, Nedim P. Vogt and Anton K. Schnyder (ed.), Basle/Frankfurt, Helbing & Lichtenhahn, 1996, 1415.

[21]    On the conditions of enforceability of consent awards, notably the elements such an award must include, Gino Lörcher, “Enforceability of Agreed Awards in Foreign Jurisdictions”, 17 Arb. Int’l. 275 (2001).

[22]    The notion of social sub-system refers to that of Günther Teubner, for instance in Günther Teubner, “’Global Bukowina’: Legal Pluralism in the World Society”, Global Law Without a State, Günther Teubner (ed.), Dartmouth, Aldershot, 1997, 3-30.

[23]    Or ‘self-execution’: see Schultz, Kaufmann-Kohler, Langer, Bonnet, “Online Dispute Resolution: The State of the Art and the Issues”, 72 (“Self-execution under the UDRP”).

[24]    Self-enforcement increases the effectiveness of the law that this form of ODR applies, thereby making this form of ODR an important instrument of legal pluralism and self-regulation, Thomas Schultz, “Online dispute resolution (ODR): résolution des litiges et ius numericum”, 48 Revue Interdisciplinaire d’Études Juridiques 153 (2002). In other words, self-enforcement certainly strongly increases the “shadow of the law”, alluding to the law which is referred to—either directly applied or looming in the background as a threat—, see Ethan Katsh, Janet Rifkin and Alan Gaitenby, “E-Commerce, E-Dispute, and E-Dispute Resolution: In the Shadow of “eBay Law””, 15 Ohio St. J. on Disp. Resol. 705 (2000), 728 (“as we encountered disputants and observed them as they participated in our [ODR] process, … we became persuade that disputants were, indeed, participating as if they were "in the shadow of the law." The law whose shadow was affecting them, however, was eBay's law rather than the shadow of any other law. … EBay was important to them, and eBay ran its site in such a way that a user's eBay future could be affected by disputes that arose. If they ignored eBay law, they did so at some risk to their future online life and even to their economic wellbeing”).

[25]    See for instance David E. Sorkin, “Judicial Review of ICANN Domain Name Dispute Decisions”, 18 Computer & High Tech. L.J. 35 (2001), 47 (“in fact a UDRP decision lacks formal legal status, unlike a court judgment or an arbitration award”) and 55 (“Judicial deference to UDRP panel decisions invites serious substantive and procedural objections“ and “De novo review is consistent with the purposes of the UDRP and is needed to ensure fairness and due process”). See also Schultz, Kaufmann-Kohler, Langer, Bonnet, “Online Dispute Resolution: The State of the Art and the Issues”, 72-73.

[26]    On the importance of communicating the availability of ODR services, for instance Janice Nadler, “Electronically-Mediated Dispute Resolution and E-Commerce”, 17 Negotiation Journal 333 (October 2001).

[27]    This is an argument that is recurrently expressed against non-binding offline arbitration. For instance Daniel E. Murray, “A Potpourri of Recent Federal Arbitration Cases Involving Domestic and International Arbitration” 13 BYU J. Pub. L. 293, 349 (“It is submitted that the addition of a non-binding additional step would not add to the efficient determination of the ultimate question. It would simply add additional delay and expense to the process”). Also Karen O'Connor, “Civil Justice Reform and Prospects for Change”, 59 Brooklyn L. Rev. 917, 928, (“This kind of compulsory, non-binding arbitration simply adds another layer onto the legal ladder and presents another potentially costly hoop for the "have nots" to jump through in their battle against the "haves." For example, in the context of a dissolution of marriage, in which women are often in inferior bargaining positions to men, the addition of this layer of "justice" simply means that some women are more likely to run out of the money necessary to vindicate their rights fully at an earlier stage of the game”).

[28]    Such laws and conventions are for instance the English 1996 Arbitration Act (Sec. 5); the US Federal Arbitration Act (Sec. 2); the Hong Kong 2000 Arbitration Ordinance (Sec. 2AC); the German 1998 Arbitration Act (§ 1031 (5) ZPO); the China 1995 Arbitration Law (Art. 16); the Belgian Code judiciaire (Art. 1667); the Dutch Arbitration Act (Art. 1021); the Swiss Private International Law Act (Art. 178 (1)); the UNCITRAL Model Law on International Commercial Arbitration (Art. 7 (2)); and the New York Convention (Art II (2)). These texts do generally not require a signature and I will therefore leave electronic signatures aside. Swedish and French arbitration laws do not require the agreement to be in writing (for Sweden: Klaus Peter Berger, “The Arbitration Agreement under the Swedish 1999 Arbitration Act and the German 1998 Arbitration Act”, 17 Arb. Int’l. 389 (2001), 395. For France, Art. 1443, 1449, and 1495 of the French Nouveau code de procédure civile; Matthieu de Boisséson, Le droit français de l'arbitrage interne et international, Paris, Gide Loyrette Nouel, 1990, 477-479.

[29]    Gabrielle Kaufmann-Kohler, “Arbitration agreements in online business transactions”, Law of International Business and Dispute Settlement in the 21st Century”, Liber Amicorum Karl-Heinz Böckstiegel, R. Briner, L.Y. Fortier, K.P. Berger, and J. Bredow (ed.), Cologne, Heymanns, 2001, 358, with extensive references.

[30]    Ibid., 361 (advocating an evolutive interpretation). The UNCITRAL Working Group II (Arbitration and Conciliation) made a declaration, after its 34th session in May 2001, regarding the interpretation of the New York Convention (NYC), recommending that its Art. II (2) should be interpreted to include a wording inspired by the UNCITRAL Model Law (Art. 7, currently under revision to alleviate any ambiguity regarding electronic records) (A/CN.9/487, § 63). In June 2002, the Working Group considered that an interpretative instrument might not be sufficient to deal with Art. 2 because it would have no binding legal effect and would thereby lead to disharmony. As the difficulties attendant upon amendment of the NYC on the development of a protocol had already been considered at earlier sessions, the Working Group acknowledged that it had not yet reached a consensus. The Working Group decided to re-examine the variety of options available (A/CN.9/508). These are an amending protocol, and interpretative instrument, and having recourse to the more-favorable-rights provision of Art. VII NYC. The discussions regarding the writing requirement have been postponed until the 38th session of the Working Group, in 2003 (A/CN.9/WG.II/WP.120, § 6).

[31]    This is comprehensively analyzed in Kaufmann-Kohler, “Arbitration agreements”, 364-368.

[32]    For a discussion of the notion of “permanent medium”, Marie Demoulin, “La notion de “support durable” dans les contrats à distance: une contrefaçon de l’écrit?”, 2000 Revue Européenne de Droit de la Consommation 361, 369 (discussing the “permanent” character of paper, CD-ROMs, floppy disks, emails, and web pages).

[33]    For instance, who should store the information? If it is the parties, there is a risk of manipulation, Pietro G. Fringuelli and Matthias Wallhäuser, “Formfordernisse beim Vertragsschluss im Internet“, 1999 Computer und Recht 93, 99.

[34]    See “Purdue Team Develops Watermark To Protect Electronic Documents”, Science Daily, 4/27/2001,

[35]    This has shortly been put into context, from the technical point of view of the non-repudiation of electronic messages, in Thomas Schultz, Vincent Bonnet, Karima Boudaoud, Gabrielle Kaufmann-Kohler, Jürgen Harms, and Dirk Langer “Electronic Communication Issues Related to Online Dispute Resolution Systems”, Proc. WWW2002 – The Eleventh International World Wide Web Conference – Alternate Track CFP: Web Engineering, Honolulu, Hawaii, conference on 7-11, May, 2002,

[36]    For a discussion of the evidential value of electronic records as arbitration agreement, see Julia Hörnle, “Online Dispute Resolution”, Bernstein's Handbook of Arbitration and Dispute Resolution Practice, John Tackaberry and Arthur Marriott (ed.), 4th ed., Sweet & Maxwell, Publication forthcoming (expected December 2002). Solutions such as burning a CD that cannot subsequently be altered is not convincing, because the document can still be forged before it is burnt on the CD.

[37]    Richard Hill, “On-line Arbitration: Issues and Solutions”, 15 Arb. Int’l. 199.

[38]    Only in Germany does the agreement to arbitrate require an additional requirement of form if the agreement involves a consumer: a signature, which can be electronic, in a separate document (§ 1031 (5) German ZPO).

[39]    Eric Loquin, “L'arbitrage des litiges du droit de la consommation”, Vers un code européen de la consommation, F. Osman (ed.), Brussels, 1998, 366; Richard Hill, “The Internet, Electronic Commerce and Dispute Resolution: Comments”, 14 J. Int’l Arbit. 103 (1997), 108. Comparing US and European law: Dirk Langer, “Arbitration and Consumer Law in Europe and the United States: Convergences and Divergences”, Latin American Journal of Mediation and Arbitration, publication forthcoming 2002,

[40]    In the case of Brower v. Gateway 2000, Inc., 676 N.Y.S. 2d 569, 1998 N.Y. App. Div. Lexis 8872, heard in the Supreme Court of New York in 1998, which concerned a consumer item of $2000-$3000, the court held unconscionable a clause mandating ICC arbitration, since the costs of the deposit for arbitration would exceed the value of the transaction itself. See also Linda Alle-Murphy, “Are Compulsory Arbitration Clauses in Consumer Contracts Enforceable? A Contractual Analysis”, 75 Temple L. Rev. 125 (Spring 2002), 158. For an overview of US practice, Joseph T. McLaughlin, “Arbitrability: Current Trend in The United States”, 12 Arb. Int’l., 123 (1996).

[41]    Veijo Heiskanen, “Dispute Resolution in International Electronic Commerce”, 16 J. Int’l. Arbit. 29 (1999), 31 (“many national consumer protection laws declare dispute arising out of transaction involving consumers as non arbitrable”).

[42]    Directive 93/13/EEC on unfair terms in consumer contracts, OJ L 095, 04.21.1993, 29-34.

[43]    This conjecture is tentatively raised in Dirk Langer, Verträge mit Privatkunden im Internet, publication forthcoming autumn 2002, introduction to Vierter Teil: Die Sicherung des Zugangs zum Recht bei grenzüberschreitenden Verträgen.

[44]    In France, Philippe Delebecque, “Arbitrage et droit de la consommation”, 104 Droit & Patrimoine 46 (2002), 50. See also Thomas Clay, “Nouvelles perspectives en matière d’arbitrage. Ouverture”, 104 Droit & Patrimoine 40 (2002), 42. In the UK, Valentina Allotti, “La clausole arbitrali nei contratti con i consumatori: l’esperienza ingles”, 8 Rivista dellarbitrato 360 (1998). In Italy, Vincenzo Vigoriti, “ Note su arbitrato e tutela di interessi minori nell'esperienza italiana e comparativa ”, 8 Rivista dell’arbitrato 445 (1998). In Germany, Norbert Reich, “Zur Wirksamkeit von Schiedsklauseln bei grenzüberschreitenden Börsentermingeschäften, Entscheidung des Court of Appeal vom 12. Juli 1996 un des Oberlandesgerichts Düsseldorf vom 8. März, 1996,” 1998 Zeitschrift für europäisches Privatrecht 981, 990. In Portugal, Isabel M. Cabeçadas, “Le centre d’arbitrage des litiges de la consommation de Lisbonne, 1999 Revue Européenne de Droit de la Consommation 391. In Spain, Manuel-Angel López Sánchez and Marta Orero Núñez, “Le système espagnol d’arbitrage des litiges de consommation”, 1996 Revue Européenne de Droit de la Consommation 120.

[45]    According to the ABA Task Force “ODR is not used to any meaningful degree in the B2B market segment since the parties have made other arrangements for the settlement of disputes between them and disputes among them are rare in any case”: ABA Task Force on Electronic Commerce and Alternative Dispute Resolution, “Addressing Disputes in Electronic Commerce: Recommendations and Report”, draft March 2002, p. 15.

[46]    Laurent A. Niddam, “Unilateral Arbitration Clauses in Commercial Arbitration”, 5 Arb. Disp. Res. L. J. 147 (1996). William W. Park, “Making Sense of Financial Arbitration”, 2000 ICC Bull. (Special suppl. on arbitration, finance and insurance) 7, 12.

[47]    Conducting the arbitration proceedings online is not as such a legal obstacle. Either the parties have no right to an oral hearing (for instance Sect. 34 English 1996 Arbitration Act) or this right can be waived (for instance Art. 24 (1) UNCITRAL Model Law). This is more clearly discussed in Hörnle, “Online Dispute Resolution”.

[48]    On the properties of an architecture of security Lessig, Code and Other Laws of Cyberspace, 39-42. On online evidence, Chris Reed, “Legally Binding Electronic Documents: Digital Signatures and Authentication”, 35 Int'l Law. 89 (Spring 2001). On some technical implications this issue has, Vincent Bonnet, Karima Boudaoud, Michael Gagnebin, Jürgen Harms and Thomas Schultz, “Online Dispute Resolution Systems as Web Services”, Proc. Hewlett-Packard OpenView University Association Workshop, held on videoconference, June 11-13 2002,

[49]    The issue is slightly different in B2B and in B2C disputes, as in B2B the parties are likely to request a physical hearing anyway: as Mirèze Philippe, “Where is Everyone Going With ODR?”, 2002 International Business Law Journal 167, 176: “B2B procedures are hybrid because it is unlikely that parties or their lawyers accept not to be personally heard by the arbitrators. […] B2B disputes are unlikely to be handled solely online, because transmitting documents in electronic format is one thing, and pleading the case or hearing the witnesses is another thing”.

[50]    Discussed in sports arbitration: Kaufmann-Kohler and Peter, “Formula One Racing”, 185.

[51]    The Bolero electronic bill of lading provides exactly such a central registry, called the ‘Title Registry Record’, which “logs and tracks all transactions centrally”: <>.

[52]    The writing requirement also exists under the UNCITRAL Model Law (Art. 31). Cf. the English Arbitration Act, which requires a writing unless the parties have agreed otherwise (Sec. 52). See also Isabelle Manevy, “Online dispute resolution: what future”, LL.M. Master thesis, Uni. Paris I, June 2002, 43,

[53]    van den Berg, The New York Convention of 1958, 250. The same provision exists in French law (Art. 1499 of the French Nouveau code de procédure civile), Jean Robert, L'arbitrage: droit interne, droit international privé, Paris, Dalloz, 1993, 284, § 319.

[54]    The UNCITRAL Working Group on Electronic Commerce, mentioned this issue in its 39th session, 11-15 March 2002, § 152: “Difficulties for the use of electronic communications may result, in particular, from the requirement, in article IV, paragraph 1, that, in order to obtain recognition and enforcement of an arbitral award, the moving party must supply: “(a) the duly authenticated original award or a duly certified copy thereof”; and “(b) the original agreement referred to in article II or a duly certified copy thereof”. In view of the growing interest in online dispute settlement mechanisms, sub-paragraph (a) of this provision may be a source of legal uncertainty, in particular in States that have not enacted legislation implementing the Model Law on Electronic Commerce, in particular its article 8, or do not other-wise provide for the functional equivalence between data messages and paper-based originals.” The current trend seems to be in favor of additional material, such as an additional protocol.

[55]    Jasna Arsic, “International Commercial Arbitration on the Internet - Has the Future Come Too Early?”, 14 J. Int’l. Arbit. 209 (1997), 217 (“Though nowhere specifically required by the New York Convention, it is questionable whether an arbitral award itself—made on the Internet and written in an electronic version with the digital signatures of arbitrator—would qualify as an original copy of the arbitral award. One practical solution to this problem would be to send the printed version of the arbitral award to the arbitrators to sign it, or to use "trusted third party" services to confine that the digital signatures are those of the arbitrators”). Michael E. Schneider and Christopher Kuner, “Dispute Resolution in International Electronic Commerce”, 14 J. Int’l Arbit. 5 (1997), 24 (“the parties to an online arbitration are well advised to require the arbitrators to issue a hard copy of the award, in order to comply with the requirements of form to make it “binding on the parties”).

[56]    Cf. Arnold Vahrenwald, “Out-of-Court Dispute Settlement Systems for E-Commerce, Report on Legal Issues”, Part IV, “Arbitration”, Report of the JRC to the EU DG Information Society, Ispra, 2000, 112,

[57]    Ethan Katsh and Janet Rifkin, Online Dispute Resolution, Resolving Conflicts in Cyberspace, San Francisco, Jossey-Bass, 2001, 108-109 (“While a loser in such a process could still go to court, it is likely that the litigation option will not be exercised very often if the losing party senses that they have obtained a fair hearing and that their position was not as persuasive as they might have thought it was”). See also, focused on offline non-binding arbitration, Rau, “Contracting Out of the Arbitration Act”, 242 (“the parties may think of a “trial run” of their case, ending in a prediction by a neutral expert, may cause the more recalcitrant among them to reassess their own partisan estimates of the likely outcome of adjudication”).

[58]    Thomas Schultz, “Online Dispute Resolution: An Overview and Selected Issues”, UNECE Forum on ODR, Geneva, June 6-7, 2002, Publication of the United Nations, publication forthcoming, (3.4:Technological architecture of ODR systems), very briefly summarized in Thomas Schultz, “The architecture of ODR systems as their best promoter”, World Of Arbitration Newsletter, Issue 2, 2002, Cf., on online mediation—keep in mind that evaluative mediation resembles non-binding arbitration— Joel B. Eisen, “Are We Ready for Mediation in Cyberspace?”, 1998 B.Y.U. L. Rev. 1305, 1310 (doubting the feasibility of online mediation).

[59]    On the control of a valuable resource as a source of private regulation, Henry H. Perritt, “Towards a Hybrid Regulatory Scheme for the Internet”, 2001 U. Chi. Legal F. 215, 237. For the application of this concept to ODR, concluding that the ODR providers may become an important actor of a new ius numericum, see Schultz, “Online dispute resolution (ODR) : résolution des litiges et ius numericum”, 196.

[60]    Some ODR providers deliver trustmarks. They are MARS,, SquareTrade, TRUSTe,, Web Trader and Word&Bond.

[61]    Definitions are provided by the ABA Task Force, “Recommendations and Report”, 45, and the Joint Research Commission of the European Commission, “E-Commerce and Consumer Protection. A Survey of Codes of Practice and Certification Processes”, report by Guido Nannariello, Institute for the Protection and Security of the Citizen, Cybersecurity Sector, EUR 19932 EN, 2001, p. vii. On the legal nature of certificates, Olivier Cachard, La régulation internationale du marché électronique, Paris, L.G.D.J., 2002.

[62]    Exactly such a clause is planned to be implemented in the Online Confidence Trust Seal—the trustmark program of the ODR project of the Eurochambers (information provided by Vincent Tilman, coordinator of the Online Confidence Project, during an interview in Brussels, on June 19th, 2002). On the conclusions drawn from the ECODIR experience, Cruquenaire and Patoul, “Le développement des modes alternatifs de règlement des litiges de consommation : Quelques réflexions inspirées par l’expérience ECODIR”, § 40-41 (“it seems desirable to connect ADR to a labeling system in order to provide a means to ensure the implementation of the agreements or recommendations that are rendered by ECODIR. A label would allow to contractually force the traders to execute the solutions found on the ECODIR platform”).

[63]    All registrars of domain names in .com, .net and .org are contractually linked to ICANN, which is in turn linked to the ODR providers applying the UDRP. The registrars are contractually bound to enforce the decisions of these ODR providers: this is an example of networking with the controlling entities.

[64]    On the current lack of trust and confidence in electronic commerce and its reasons, B. Brun, “Nature et impacts juridiques de la certification dans le commerce électronique sur Internet”, Lex Electronica, vol. 7, n°1, été 2001, <>

[65]    Research conducted by BBBOnline reported that 84% of web users declare that they would be reassured by a certificate. This research has been reported in T. Trompette, “Une nouvelle mission : la certification des sites Web de commerce électronique”, Les Cahiers de l’Audit, vol. 4, 1999, p. 34. Other empirical research has shown that “[t]he presence of credit card symbols do little to communicate trustworthiness, even though they’re universally recognized by consumers. In contrast, Web-based “security brand”, seals of approval, such as VeriSign, when recognized, do communicate trustworthiness”: Cheskin Research et Studio Archetype/Sapient, “Commerce Trust Study”, January 1999, <>.

[66]    For instance Pierre Trudel France Abran, Karim Benyekhlef and Sophie Hein, Droit du Cyberespace, Montreal, Les Éditions Thémis, 1997, p. 3-46.

[67]    For instance Canada: see Industrie Canada, “Magasiner dans Internet. Renseignez-vous”, 8 November 1999, and Bureau de la consommation, Industrie Canada, Votre commerce dans Internet. Gagner la confiance des consommateurs, 8 November 1999, SSGF/ca01186f.html#Certifying.

[68]    Katsh, Rifkin, Gaitenby, “E-Commerce, E-Dispute, and E-Dispute Resolution”, 731 (“A somewhat less obvious eBay law or legal process concerns the power of exclusion, a power that, in the context of eBay, is a power over existence. This may not be a power that often is exercised, but for it to have effect, it is less necessary that it be used than that buyers and sellers are aware that it could be used. In our pilot project, where mediation was the sole process used, participants had no reason to fear being evicted from the marketplace. A marketplace could, however, rely on an arbitration process rather than mediation and use the threat of exclusion as the mechanism for enforcing the terms of the ruling”).

[69]    Example of escrow services are (, which offers specific services for eBay users, and (

[70]    See for instance’s FAQ: “What if there is a disagreement during the transaction? What is dispute resolution? If the Buyer and Seller haven't resolved the dispute within 60 days, the matter is sent to dispute resolution - an arbitration process administered by the American Arbitration Association in accordance with the provisions of its Commercial Arbitration Rules” (

[71]    Norbert Wühler, “The United Nations Compensation Commission: A new contribution to the process of international claims resolution”, 2 J. Int’l Econ. L. 249 (1999), 250 (“Compensation is paid to successful claimants from a special fund that receives a percentage of the proceeds from sales of Iraqi oil”) 269 (“the sale of oil by Iraq is the source of income for the Compensation Fund, with 30 percent of the revenue from such sales being deposited in the Fund for the payment of compensation and for operating costs of the Commission”).

[72]    In August 2000, a German Act came into force, creating a German Foundation entitled “Remembrance, Responsibility and Future”, to provide financial compensation to former slave and forced laborers and other victims of Nazi injustice. The funds for the Foundation (DEM10 billion) were made available by the German Government and German companies. Under the Act, the International Migration Office (the “dispute resolution provider”) has been allocated DEM 740 million to pay eligible persons filing claims. Many claims are made by email and there are no physical hearings, therefore it is considered to operate partly online. See

[73]    Trusted Shops, for instance, offers a similar service: “If the goods are not delivered, Trusted Shops will take action immediately after you have registered your complaint. If the dispute cannot be resolved, Gerling will refund your advance payment upon approval of your claim, and “Trusted Shops with money-back guarantee from Gerling insures you when shopping online against financial loss due to non-delivery or return of goods”,

[74]    See EU Directive on distance contracts (Directive 97/7/EC of the European Parliament and of the Council of 20 May 1997 on the protection of consumers in respect of distance contracts, OJ L 144, June 4, 1997, 19-27), Art. 8: “Payment by card: Member States shall ensure that appropriate measures exist to allow a consumer: - to request cancellation of a payment where fraudulent use has been made of his payment card in connection with distance contracts covered by this Directive, - in the event of fraudulent use, to be recredited with the sums paid or have them returned”. On European laws on credit card risk allocation, see Rufus Pichler, “Kreditkartenzahlung im Internet”, 1998 Neue Juristische Wochenschrift 3234.

[75]    For a comparison between US and European policies regarding credit card chargebacks, Rufus Pichler, “Finality of Credit Card Payments and Consumer Confidence—Different Approaches in the United States and in Europe”, 5 Electronic Payment Systems Observatory Newsletter (2001),

[76]    For the US, see the Truth-in-Lending Act, Title I of the Consumer Credit Protection Act, 15 U.S.C. §§ 1601-1667 and the Federal Reserve Board’s implementing regulation Z, 12 C.F.R. part 226 (1999). For a discussion of US credit card risk allocation, Jane Kaufmann Winn, “Clash of the Titans: Regulating the Competition Between Established and Emerging Electronic Payment Systems”, 13 Berkeley Tech L. J. 675 (1999).

[77]    On credit card chargebacks as “a general enforcement model”, see Rufus Pichler, “Trust and Reliance—Enforcement and Compliance: Enhancing Consumer Confidence in the Electronic Marketplace”, Stanford University, May 2000, p. 155-158, Considering credit card chargebacks as an alternative dispute resolution mechanism, Henry H. Perritt, “Dispute Resolution in Cyberspace: Demand for New Forms of ADR”, 15 Ohio St. J. on Disp. Resol. 675 (2000), 693.

[78]    Art 4(k) of the UDRP.

[79]    The control of this database carries an important control of the Internet, Neil Weinstock Netanel, “Cyberspace Self-Governance: A Skeptical View from Liberal Democratic Theory”, 88 Calif. L. Rev. 395 (2000), 485 (“ICANN's future power should not be underestimated. No one can establish a publicly accessible web site without an Internet domain name. Accordingly, if ICANN should decide that domain name registrars may (or must) deny registration unless the applicant forswears certain sorts of expression, meets specified criteria of “good standing”, or pays a substantial fee, then those who fail to do so will effectively have no presence on the web”).

[80]    In order to obtain a domain name in .com, .org or .net, the prospective domain name holder must make a contract with a registrar approved by ICANN—at least when the website is not meant for techies, but techies a special. ICANN approves only registrars that commit to enforce UDRP decisions.

[81]    See for instance Elizabeth G. Thornburg, “Going private: Technology, Due Process, and Internet Dispute Resolution”, 34 U.C. Davis L. Rev. 151 (2000), 197 (“If the domain name holder who has lost the ICANN proceeding cannot afford to hire an attorney to draft a complaint and to pay the filing and service fees for a lawsuit, the domain name is gone”.

[82]    Lawrence Lessig, “The Law of the Horse: What Cyberlaw Might Teach”, 113 Harv. L. Rev. 501 (1999), 530-531 (“The dissimilarity is this: with every enforced contract - with every agreement that subsequently calls upon an enforcer to carry out the terms of that agreement - there is a judgment made by the enforcer about whether this obligation should be enforced. In the main, these judgments are made by a court. […] When the code enforces agreements, however, or when the code carries out a self-imposed constraint, these public values do not necessarily enter into the mix.”). See also Lessig, Code and Other Laws of Cyberspace, 136 (“The ultimate power of a contract is a decision by a court–to enforce the contract of not. […] The same is not true of code.”).

[83]    Elizabeth Longworth, “The Possibilities for Legal Framework for Cyberspace—including a New Zealand Perspective”, The International Dimensions of Cyberspace Law, Teresa Fuentes-Camacho (ed.), Paris, UNESCO Publishing / Ashgate, 2000, 9, 25-26.

[84]    Lessig, Code and Other Laws of Cyberspace, 88-89.

[85]    Perritt, “Towards a Hybrid Regulatory Scheme for the Internet”.

[86]    One of Perritt’s examples is ICANN. The authority of ICANN over domain name holders, Perritt argues, does not derive so much from its legal status than from the control by ICANN of the resource that is valuable to the domain name holders: the databases that converts domain names into IP addresses. Without this database—controlled by ICANN—the access of the WebPages of a domain name to the Internet is largely hindered, if not completely denied. Ibid., 237, 238 (“By refusing to list a domain name in authoritative domain name servers, ICANN and domain name registrars can deprive those subject to their regulatory authority of access to the internet). Now recall that ICANN approves providers of dispute resolution services and enforces their decisions without any review.

[87]    H.L.A. Hart, The Concept of Law, 2nd edition, Oxford, Oxford University Press, 1994, 6-7, 18-25.

[88]    Katsh, Rifkin, Gaitenby, “E-Commerce, E-Dispute, and E-Dispute Resolution: In the Shadow of “eBay Law””, 731.

Top | Home

© 2002 Center for Information Technology and Dispute Resolution