Detailed Explanation of Decentralized Arbitration Protocols: Kleros, Aragon, Jur

Source: DAOrayaki

Authors: Yann Aouidef, Federico Ast, and BruDeffains

Compilation: DoraFactory

In 1958, the New York Convention was signed to provide a critical governance infrastructure to address large-scale trade and investment disputes involving governments and multinational corporations. However, with the rise of the internet and digital commerce, new disputes have emerged that differ in nature from the millions of dollars in cases established under the New York Convention framework. In this new economic reality, there are numerous small international claims of a few thousand dollars, such as software development contracts, remote team crowdfunding activities that fail to deliver on promises, or online fraud (Nappert and Ast, 2020).

As the father of online dispute resolution, Ethan Katsh, stated: technology generates disputes faster than technology can resolve them (Katsh and Rabinovich-Einy, 2019).

Traditional dispute resolution systems (e.g., state courts, international arbitration) seem inadequate to address the large volume of disputes arising from online transactions. It is estimated that 3% to 5% of online transactions end in disputes (Mania, 2015). The cross-border nature of e-commerce and the demand for shorter and cheaper judicial processes provide reason to expect rapid growth in the online dispute resolution industry (Schultz, 2002).

Although ODR services have existed since the 1990s, the industry has still failed to reach the early anticipated growth potential (Deffains and Gabuthy, 2006). With the development of projects operating on blockchain and smart contracts, decentralized arbitration has emerged. This promises a novel way to provide fast and affordable dispute resolution processes for new small claims in the digital economy, injecting new vitality into the industry.

Decentralized arbitration is the result of the convergence of online dispute resolution, blockchain, international arbitration, and mechanism design. This paper primarily conducts a descriptive and comparative analysis of three projects that play pioneering roles in the decentralized justice industry: Kleros, Aragon, and Jur.

1. Decentralized Justice

Ronald Coase argued that under certain conditions, cooperation among agents can yield positive effects in terms of economic efficiency and social capital (Coase, 1937). The capabilities of blockchain facilitate value transfer and self-executing agreements, reducing transaction costs. The characteristics of blockchain—disintermediation, immutability, and transparency—enable widespread applications across various sectors, including e-commerce, finance, insurance, healthcare, and social media. However, further development is needed to realize these predictions.

Smart contracts are "smart" enough to self-execute as written in the code. However, they cannot resolve situations where parties interpret terms differently. Litigation between parties slows down the speed and automation of transaction execution, leading to high legal uncertainty and high transaction costs (Coase, 1960), thus losing all the benefits created by smart contracts. Additionally, an analysis was conducted using the Dispute Resolution Possibility Frontier (DRPF) framework, which maps and describes four institutional possibilities (private ordering, arbitration, courts, and regulatory state) in terms of disorder costs and dictatorship costs to understand how existing dispute resolution mechanisms help resolve smart contract disputes (Allen et al., 2019).

Decentralized judicial platforms are a form of "digital courts" supported by blockchain technology, aimed at resolving disputes through crowdsourced jurors under economic incentives to provide fair rulings (Ast and Dimov, 2018). The rules of these platforms are encoded in smart contracts operating on the blockchain. Decentralized judicial platforms aim to provide a method for resolving the inherent interpretative issues of smart contracts, thereby reducing transaction costs and enabling many decentralized applications built on blockchain to thrive.

From this perspective, the emergence of decentralized justice can be interpreted as an effective institutional response to the economic problem of coordination among agents to reduce transaction costs (Deffains and Gabuthy, 2006). For example, modern e-commerce has introduced a new category of disputes: small-scale cross-border disputes. Since traditional dispute resolution channels (including courts) are unsuitable for this new category, blockchain-based "decentralized justice" platforms offer a viable alternative. Some believe that decentralized justice can achieve more nuanced crypto solutions and generate greater certainty in the process (Kaal and Calcaterra, 2018).

It is decentralized because the process is peer-driven, built on blockchain technology, and cannot be controlled by any single agent. It is fair in terms of providing just resolutions for the cases submitted to it. Due to its innovative mechanism design, decentralized judicial systems have the potential to offer a fast, cheap, and fair dispute resolution solution. This solution will provide an executable path for claims in e-commerce, freelancing, crowdfunding, and other digital economy cases or more traditional disputes.

2. Key Players in Decentralized Justice

As of early 2020, the decentralized justice industry was represented by three key projects: Kleros, Aragon, and Jur.

Kleros was founded in May 2017 by Federico Ast and Clément Lesaege (Kleros website, 2020). The development of the Kleros protocol is implemented by the Société Coopérative d'Intérêt Collectif (SCIC) Coopérative Kleros, established in France. Kleros launched on the Ethereum blockchain in July 2018, becoming the first operational decentralized judicial platform and the most widely used platform at the time of writing. As of November 2020, nearly 500 disputes had been resolved, with about 400 users participating as jurors on the platform. This has generated approximately $123,000 in arbitration fees paid to jurors.

Coopérative Kleros follows a hybrid strategy targeting both native cases from the blockchain industry and mainstream use cases where traditional participants from the ODR industry are already active. For blockchain industry applications, the company's goals include dispute resolution for transactions, registration of token management (using distributed jurors and economic incentives for compliance verification), and oracle dispute resolution. For mainstream applications, Coopérative Kleros facilitates the construction of solutions by so-called "layer two companies" (Kleros Blog, 2020).

Aragon was founded in February 2017 by Luis Cuende and Jorge Izquierdo in Spain and is now incorporated into the Aragon Association, a nonprofit entity based in Switzerland (Aragon Wiki, 2020). The vision of the Aragon project is to provide users with software tools to create decentralized autonomous organizations (DAOs) (Aragon website, 2020). In November 2019, Aragon launched its decentralized court, whose mechanism design was primarily inspired by Kleros (Aragon, 2020). According to Aragon's data, the Supreme Court has 239 Aragon Network jurors (2020), but the number of resolved disputes is not shown. In the Aragon system, the court service department focuses more on providing arbitration services for DAOs within the Aragon ecosystem. The main use case envisioned in the Aragon white paper (2020) is to resolve disputes over voting proposals that may contradict the organization's "constitution" in supreme resolutions.

Jur was established in October 2017 by Alessandro Palombo and Giotto De Filippi under Swiss law as a "Société Civile/Société Commerciale." At the time of writing, Jur had not yet released a working product. The project's white paper claims that the system will cover various cases through three different courts: the court layer (similar to traditional ODR arbitration systems for handling higher-value disputes, claiming to produce legally binding rulings), the open layer (more similar to Kleros's system, with decision-making logic based on collective wisdom), and the community layer (a private court defined by specific rules set by the creators). Jur's strategy appears to focus on enterprise use cases. Initially built on Ethereum, it shifted to the VeChain blockchain in July 2018, attempting to concentrate on the enterprise sector (VeChain, 2020). The following figure summarizes some key elements of these projects.

3. Dimensions of Decentralized Justice

Decentralized justice projects make different choices regarding many dimensions of mechanism design. The similarities and differences between projects include the following dimensions:

Complexity of Cases

Different mechanism choices affect the types of disputes the platform can handle. Currently, all platforms can resolve binary situations where a decision must be made between two options. Some platforms claim to be able to handle non-binary cases.

Selection of Jurors

Different platforms have made different choices in the process of selecting jurors, particularly regarding the possibility of reviewing jurors based on specific criteria such as skills. Kleros and Aragon do not provide a mechanism for selecting jurors based on specific criteria (any user holding tokens can become a juror), while Jur allows for juror review.

Impact of Reputation Enhancement

Different platforms have different approaches to juror reputation. Kleros and Aragon rely solely on symbolically created crypto-economic incentives, while Jur takes juror reputation into account.

Appeal Mechanisms

Different platforms have made different choices regarding the possibility of users appealing decisions. Some platforms allow appeals (Kleros and Aragon), while others do not (Jur).

Governance Models

In Kleros and Jur, platform tokens are used for governance (e.g., deciding on software upgrades). In Aragon, a different token is used.

Subscription Fees

In Aragon, parties must pay a regular subscription fee to have the right to use the court's powers. In Kleros and Jur, fees are only paid when a dispute arises.

Trustless Behavior

Kleros and Aragon are built on fully decentralized blockchains that require no trust, allowing any user to join as a node. Jur is built on VeChain, which employs a proof-of-authority consensus algorithm with trusted nodes.

Different mechanism design choices require different trade-offs. The primary economic mechanism currently in use is the Schelling point, which in game theory is the default solution people choose in the absence of communication. Kleros, Aragon, and Jur are utilizing this mechanism. The idea of using it in the blockchain ecosystem was first proposed by Vitalik Buterin, one of the founders of the Ethereum blockchain, as many blockchain-based systems involve voting, which can be implicitly or explicitly seen in Schelling point-based games (essentially for decision-making in DAOs).

The theory behind the Schelling point is that if everyone wants others to vote honestly, their motivation is also to vote honestly to comply with the majority, which is why people can expect others to vote honestly; this is a self-reinforcing Nash equilibrium (Buterin, 2015). Nevertheless, social choice functions (belonging to the multi-choice Schelling game) also have some limitations:

• They miss some expected properties, such as clone independence. For example, let’s assume that among all the options provided to jurors, there are only two choices that allow one party to delay improvements to the subject of the transaction (e.g., developing a website). Suppose these two delay options differ only in that they do not allow for a clear definition of specific delays (e.g., we know we need more delay, but we do not know how much more time is needed to improve the website). In this case, although more delay is the better solution, jurors may deviate from the delay choice to avoid the risk of choosing the wrong delay. Thus, clone independence is an expected property to avoid these deviations from the optimal solution for jurors.

• They suffer from pre-revelation attacks, where jurors can decide to collude and reveal their votes. Many Anti-Pre-Revelation game strategies have been proposed (Buterin, 2015).

• They suffer from p+ε attacks, which are bribery attacks where an agent attempts to corrupt a juror's decision by promising (with variable credibility) to pay those who vote for certain specific decisions. Additionally, many mechanisms have been proposed to avoid these attacks (Buterin, 2015).

Moreover, more broadly:

• When there are three or more choices, then every weak Pareto and IIA social welfare function (which belongs to the multi-choice Schelling game) must be dictatorial (2016).

• Under very limited assumptions about social choice functions and reward structures, there will inevitably be incentives for participants to deviate from the concept of honesty.

• The Schelling point mechanism design has limitations in resolving non-binary situations but remains the most effective choice for simpler situations (usually binary situations). While it is unclear what the "right answer" is in terms of design, it is important to maintain consistency between the design and the types of use cases the platform seeks to serve.

4. Adoption of Decentralized Judicial Systems

The decentralized justice industry is still in its early stages. As of May 2020, the total number of users on decentralized judicial platforms was less than 1,000. As applications from the blockchain ecosystem are adopted, decentralized judicial platforms provide a more efficient and trustless way of transaction security compared to government courts, international arbitration, and traditional online dispute resolution methods, which will increase the usage of decentralized judicial platforms.

Due to their innovative game-theoretic mechanism design and frictionless payment tracks, decentralized judicial systems can offer a cheaper and faster adjudication system than alternatives while ensuring complete transparency of the process.

An important question is: how will decentralized judicial platforms evolve from their current early stage into efficient providers of "justice as a service"?

To answer this question, we can conceptualize decentralized judicial systems as a market coordinating the supply and demand for resolution services. The contracting parties (e.g., two users using a decentralized judicial platform to resolve an e-commerce dispute) are the demand side of the market. They utilize a adjudication service provided by jurors through the platform's crypto-economic mechanisms. Jurors are the supply side of the market. They sell their time and skills to process relevant information and make decisions. Traditional legal and arbitration systems are costly because the institutional arrangements under which they operate often grant legal experts a monopoly on legal services: lawyers monopolize legal advice, and local judges monopolize rulings with dominion and enforcement power. This leads to a limited supply of adjudication services, resulting in high equilibrium prices.

Due to its crypto-economic mechanisms, decentralized judicial platforms can leverage the knowledge and work of individuals with specific expertise, which may not necessarily be recognized by the legal skill development system. As shown in Figure 3, as more agents participate as service providers, the supply of resolution services increases, allowing the cost per case to reach levels unattainable by traditional methods. The quality of adjudication services is not reliant on professional institutions and codes of conduct but rather on pure economic mechanisms. This does not mean that decentralized judicial platforms replace legal professionals in resolving legal claims. However, it does mean that they can address some types of disputes that legal professionals are not well-suited to handle.

Decentralized judicial platforms rely on network effects to create low-cost dispute resolution solutions. When parties use a decentralized judicial platform, this generates demand for resolution services, thereby increasing prices and incentivizing jurors to join the network. The increase in the supply of dispute resolution services drives the market adjustment mechanism towards equilibrium, lowering the costs of dispute resolution services. As each cycle progresses, more users join the network, and the increasing specialization will yield better, cheaper, and faster decisions.

In the early stages of a network effect-based business model, decentralized judicial networks face the "chicken and egg problem": which comes first, the cases to be resolved (demand) or the jurors to resolve them (supply)? If there are no cases to resolve (and thus no revenue), what is the motivation for jurors to join this network? If there are not enough jurors to resolve cases, what is the motivation for parties to use this platform? To initiate the network effect, decentralized judicial platforms face this "chicken and egg" problem.

An unresolved question is whether the structure of the decentralized judicial market will exhibit "winner-takes-all" behavior like many network effect-based digital industries. As network effects expand, it typically benefits from economies of scale, creating advantages over competitors. In the case of decentralized justice, this can manifest in the following way: a platform gaining an advantage will lead to more cases and jurors, allowing efficiency gains to bring in more cases and higher incentives for users to join the juror pool. At some point, this network may become large enough that followers find it difficult to overcome the efficiency gains.

Whether the decentralized justice industry will indeed exhibit such behavior remains to be seen. We can argue that cost efficiency (as the lowest-cost network) is not the only relevant competitive variable. Another important variable could be the network's ability to comply with certain institutional conditions (ethical or regulatory requirements). For example, although Uber successfully established a larger network of drivers and passengers, it is banned in some jurisdictions that deem it non-compliant with appropriate regulatory conditions. In these jurisdictions, other local participants may be the leaders.

Similarly, different mechanism designs adopted by various decentralized judicial platforms may be institutionally better suited for different types of use cases. If this is the case, then the industry may ultimately split into different providers. For instance, as the lowest-cost provider, Platform A may become the leader for use cases where cost is the key competitive variable. However, due to ethical or regulatory reasons, the mechanism design chosen by Platform A may not be suitable for other types of applications. These types of use cases could be served by Platform B. Although it may not be the lowest-cost provider, it could be better suited from an institutional perspective.

Here, a deeper understanding of the similarities and differences in mechanism design among decentralized justice projects can help clarify the different scenarios they may serve.

5. Conclusion

The global legal market is valued at over $1 trillion, while the blockchain market is estimated at $15 billion. As of May 2020, the total market value of decentralized judicial platforms was still below $10 million, with the total number of users remaining under 1,000. This demonstrates that the decentralized justice industry is still in its early stages. However, the number of low-value digital transactions is rapidly increasing, providing a growing market for effective and fair dispute resolution methods. For example, the value of global cross-border payments reached $25.8 billion in 2019 and is expected to reach $35.6 billion by 2022, which will increase as more transactions begin to be conducted in cryptocurrency, providing an almost frictionless experience for international payments.

In this article, we primarily defined decentralized justice, analyzed the main projects in the market, proposed some key dimensions for analyzing the industry, and discussed the different options for how the future market may grow and evolve.

The world economy is rapidly changing and requires a new institutional framework to ensure transaction security. This heralds a fundamental change in legal practice. As industry expert Richard Susskind stated, "The changes in the legal industry over the next 20 years will be greater than those of the past 200 years."

Naturally, these changes will generate significant debates from a legal perspective, such as whether dissenting judgments are considered valuable for legal development. What makes blockchain technology a powerful tool for promoting disintermediation and decentralized coordination. Collective action is primarily aimed at reaching compromises between conflicting interests and values, while blockchain technology operates through distributed consensus and exit-based conflict resolution systems. Therefore, it is important to consider how decentralized justice can promote the common good, generate shared benefits, and create a sense of collective justice through more market-based solutions.

An important part of this transition is likely to come from the fields of machine learning and legal analytics. Another part will come from the emerging decentralized justice domain that incorporates crowdsourcing and mechanism design into dispute resolution processes.