Blog 9 - Monitoring the members’ of blockchain organisations behavior.
Members interaction with common assets on public blockchains are completely transparent to all members and tools are widely available to monitor activity. Although Ostroms rule appears to be satisfied by existing characteristics of blockchains for those organizations that operate solely on a blockchain – those that have completely ceded to the crypto economy – beyond pure cryptocurrencies, it is difficult to imagine many scenarios where this is the case. That is for the crypto economy to achieve its potential, DAOs will need to interface with the real economy and real people to varying degrees and their challenge is therefore establishing and maintaining reputations of the members they will rely on to input off chain inputs. Considering the members are both responsible for the governance and development of the DAO and are the customers/suppliers of the P2P business that the DAO operates, this is a multifaceted consideration.
To provide a prism to look at what is required, let’s consider a possible DAO. Uber is a platform that matches drivers with riders and takes a 30% cut for providing the service. It also measures participant’s behavior by requesting that both drivers and riders rate the experience each time they use the service. A DAO could provide a similar service to UBER, distributing the 30% cut to members as discounted services, increased revenue or dividends. It could also copy Ubers reputation system, reporting to other members and taking away privileges based on member’s reputations. Beyond the challenge discussed in an earlier blog of ensuring that member’s identities cannot be easily recreated and therefore reputations systems circumvented, the members of such a DAO would also be responsible for contributing to the definition of the DAO through buildings its code, participating in its governance process and building its membership base. In the Uber DAO example, the members may be the coders, shareholders, employees, board of directors, marketers etc as well as the consumers of the service. To what extent would a DAO need to report and monitor member’s behavior so as other members can predict if member’s actions and opinions are aligned with their own best interests? How can this be best achieved?
Bitcoin is the oldest and most successful DAO and an open system that uses blockchain to achieve consensus on the validity of transactions. Bitcoins governance was initially managed by a single individual – Satoshi Nakamoto – and he used his relationships with potential contributors to inform his assessment of their trustworthiness before sharing and ultimately handing over authority to them. As bitcoin membership numbers grew beyond the capability of individuals to hold peer to peer relationships it increasingly began to rely on social networks such as reddit and twitter to establish the reputations of members who develop and govern it. Although this unstructured method is clearly aligned to Ostroms design principle and it is arguable that this method for establishing members’ reputations has greatly contributed to bitcoins initial success, it is also questionable whether it is sufficient for bitcoin to evolve to a state of maximum utility and value for the majority of its members. An obvious risk with this method is the possibility of members portraying one behavior on social media and doing something else entirely different in private as part of a strategy to improve their personal position at the expense of other members.
Blockchains such as Tezos that use Delegated Proof of Stake (DPOS) as an alternate to bitcoins POW consensus system somewhat mitigate this risk. Tezos issues the rights to produce blocks and vote on protocol changes to members based on their cryptocurrency holdings as a % of the total issued. This is distinct from bitcoins competitive computational market for block validation and use of social networks to reach consensus on protocol changes. Tezos’s Block producers are required to stake their personal holdings as a bond that may be forfeited if they act outside of the blockchain protocols rules and are incentivized to compete for other members’ rights. In attracting these rights, block producers associate their social identity and their blockchain public address and thus provide other members with visibility of both their portrayed identity and the actions they take. Although this is an opt in ecosystem and members may remain anonymous, it is arguable that these Tezos members who present a social identity inclusive of a record of their blockchain behavior will likely become the trusted key opinion leaders of the Tezos DAO and Ostrom’s principle has been far more optimally achieved.
Although our hypothetical Uber DAO will implement Ubers methods of monitoring and recording driver and rider’s behavior and the underlying Tezos blockchain on which it operates is optimally compliant with Ostrom’s principle, it is arguable that this is not enough for our DAO. The model for profiting from building a DAO is typically to sell tokens that represent fractional ownership, retain a percentage and drive capital appreciation into the token price by increasing its utility. Although under this scenario the ownership of shares can be traced back to a public blockchain address, there is no correlation between investors in the DAO and the block producers whose identity and reputation are established. That is, we need to establish a separate reputation for the DAO members who are participating in its governance, and if we lack a decentralized method for doing so, we will need to revert to a centralized governance model and will no longer have a decentralized organization. Our DAO becomes obviously pointless.
For our DAOs to interface with the real world we need a decentralized method of establishing and providing ongoing reputation and it seems logical that this same system is also used to establish identity. Additionally, it seems obvious that it would be far more efficient if this reputation were in itself provided by the underlying blockchain, another blockchain or a specific DAO that other DAOs could leverage - Rather than establishing identity and reputation every time. It is also evident in the attempts to build decentralized reputational system that they must depend on the incentivizing of members to actively build an identity and consent to report on their reputation and to evaluate others. Verity is/was an attempt to build an identity and reputational system as a DAO on the Ethereum blockchain, and although there may be much merit in the algorithms proposed in its operation, it is arguable that its’ fundamental flaw is a failure to achieve value in the tokens distributed as incentives and that this flaw is insurmountable for any DAO based reputational system. DREP, a yet to be launched blockchain, proposes to overcome this issues by utilizing its native cryptocurrency as the currency to incentivize its reputations markets. Although there are obvious precedents of blockchains being able to develop cryptocurrencies that have value and therefore it is possible that DREP will overcome Veritys’ weaknesses, DREP is not intending to support programmable smart contracts or DAOs and it is therefore relying on technologies that enable cross blockchain interactions, so it can provide its utility to those blockchains that require support smart contracts and DAOs. This is a risky design assumption as cross disparate blockchain technology is currently rudimentary at best, is progressing slowly, and may never achieve the levels of integration required for this approach to be successful. This leaves us with the option of the baking of identity and reputation systems into the fabric of general purpose (smart contract) blockchain platforms for use by the DAOs that it hosts, and using underlying cryptocurrency to incentivize participation in a reputations market. A possibility to realize such an option is to leverage the reputation built by the block validators in Tezos type DPOS blockchains to determine and issue reputational credits. Using an I trust A, and A trust B (the delegate), so I trust B type logic, it is conceivable that members would trust an average of delegates assessment of other members behaviors.