A person owed a few hundred pounds by a builder, a landlord or an online seller has a legal right and, very often, no sensible way of enforcing it. The classic account of the access to justice movement described successive waves of reform aimed at this problem: legal aid, representation of diffuse interests, and then a broader rethinking of procedures and forums so that ordinary disputes could be handled in proportion to their size (Cappelletti and Garth, 1978). Small claims courts belonged to that third wave. Online dispute resolution, usually shortened to ODR, is its most recent expression.

The scale of the unmet need is well documented. Surveys of legal need in many countries show that civil justice problems are common, that most people who experience them never reach a lawyer or a court, and that a substantial share simply give up (Genn, 1999; OECD and Open Society Foundations, 2019). The pattern holds across income groups and is sharpest among those with the fewest resources. Against that background, a process that can be started from a telephone, at any hour, without a lawyer and without a day off work has an obvious appeal.

Whether the appeal is borne out is the question this review addresses. It asks what the published literature says about ODR as a means of access to justice for low-value civil claims, where the evidence is firm, and where claims rest mainly on design intentions.

This is a narrative review. It draws on the principal monographs on ODR and online courts, on law review and socio-legal journal articles, and on official reports and legal instruments from England and Wales, Canada, the United States and the European Union, which are the jurisdictions most discussed in the English-language literature. Sources were selected for their influence on the debate and for the range of positions they represent. No claim is made to have identified every relevant study, and the review does not pool results. "Low-value" is used in the sense each system gives it: the small claims track in England and Wales covers most claims up to 10,000 pounds, and the European Small Claims Procedure, created by Regulation (EC) No 861/2007, now applies to cross-border claims up to 5,000 euros.

ODR began outside the courts. When strangers started trading with one another over the internet in the 1990s, neither litigation nor conventional mediation could deal with the resulting disputes at a cost that made sense, and the platforms built their own machinery. Katsh and Rifkin (2001) described the technology in these systems as a "fourth party" that does part of the work a human neutral would otherwise do: structuring communication, narrowing issues, proposing standard solutions. The marketplace systems came to handle tens of millions of disputes a year, most without human intervention (Katsh and Rabinovich-Einy, 2017).

Public bodies took notice. The European Union legislated for a consumer ODR platform linking traders, consumers and approved dispute resolution bodies (European Parliament and Council, 2013), building on a varied national landscape of consumer ombudsmen and complaint boards. The United Nations Commission on International Trade Law, unable to agree binding rules, issued non-binding technical notes describing a staged process of negotiation, facilitated settlement and a final stage (UNCITRAL, 2017). In England and Wales an advisory group of the Civil Justice Council proposed an internet-based court for claims up to 25,000 pounds with three tiers: online evaluation, online facilitation and online judges (Civil Justice Council, 2015). The review of the civil courts that followed adopted the idea in modified form and stressed that the new court should be designed for use without lawyers (Briggs, 2016).

The models now in operation differ in who runs them, whether participation is voluntary, and what happens when the parties do not agree. Table 1 summarises the main types discussed in the literature. The Civil Resolution Tribunal of British Columbia is the example most often cited of a fully public system: created by statute, it takes parties from a guided self-help tool through negotiation and facilitation to a binding decision, and its designers have described it as built around the user and not around the institution (Salter, 2017). Rabinovich-Einy and Katsh (2017) call such bodies "the new new courts", marking a second shift after the alternative dispute resolution movement of the late twentieth century.

Access to justice is not one thing. A system may make it easier to start a claim and no easier to win one, or quicker to reach an outcome and no more likely that the outcome is right. Sandefur (2019) has argued that the measure should be the just resolution of people's problems and not their consumption of legal services or of court time. The literature is easier to read if the stages of a claim are kept apart, as in Figure 1, which sets beside each stage the barrier that writers most often say an online process leaves in place.

Four gaps stand out. The first concerns non-users: evaluations report on those who complete an online process and say little about those who abandon it or never begin. The second is the counterfactual. Faster disposal is informative only if the comparison group is credible, and few published studies have one. The third is substantive accuracy, which would require independent assessment of a sample of outcomes against the governing law. The fourth is enforcement. A decision that is not paid is not a remedy, and the literature reviewed here has largely stopped at the point of decision.

There are also lessons in projects that did not last. The Dutch Rechtwijzer platform for separating couples, widely admired by reformers, was withdrawn in 2017 when it could not sustain itself financially, and the European Union has since legislated to close its consumer ODR platform after years of low use by traders. Neither episode shows that ODR fails, but both show that technical quality does not secure uptake, and that voluntary systems depend on the stronger party having a reason to take part. Future work would benefit from independent access to platform data, from designs that follow claimants beyond the decision, and from closer attention to defendants.

The literature supports a measured view. ODR lowers the cost and inconvenience of pursuing a small claim for people who are able and willing to use it, and well-designed systems can be experienced as fair. It has not been shown to reach those who are excluded from conventional justice by something other than cost, to redress the advantages of repeat players, or to deliver outcomes that follow the law. These are questions about institutions and evidence more than about technology, and they can be answered only if the bodies that run online systems open them to independent study.

To be declared by the author(s) before publication.

To be declared by the author(s) before publication.