Token-Curated Registries (TCRs) are a class of token design that gained popularity within the crypto community in 2017.
In simple terms, TCRs are incentive systems designed to produce a list. There are three types of participant:
- Applicants apply to be included on the list.
- Curators vote to admit or to reject applicants.
- Consumers access information on the curated list.
All sorts of economic incentives can be baked into this system, but at the basic level:
- Applicants submit fees along with their applications.
- Curators are token holders who may earn tokens for their work.
- Consumers access the data for free.
While this is usually how the three groups are aligned economically, alternative designs are certainly possible, where for instance applicants are rewarded by consumers.
The purpose of this essay is not to examine advanced TCR designs, as there are virtually an infinite number of possibilities. Instead, we want to
- Answer the fundamental question of: “why do you need a blockchain?”
- Review known problems with TCRs and propose practical solutions.
Why do you need a blockchain?
The first question that everyone should ask about every blockchain project is: “Why do you need a blockchain for this?” If one cannot articulate the value proposition of using a blockchain, it’s almost guaranteed to be a shitcoin project that is taking advantage of the retail hype.
TCRs are no exception. They are certainly not a magical solution to every curation problem. In this section we’ll list the reasons why a TCR may make more sense than its alternatives. Alternatives being:
- One central curator: e.g., Michelin, Forbes 30 under 30.
- One or many curators on a central platform: e.g., college admission, Reddit, Yelp.
- Free curation: e.g., Reddit, Yelp
Advantages vs. One Central Curator
- Wisdom of the crowd: TCRs are weighing machines. They leverage the idea that, under the right circumstances, the many are smarter than the few.
- Risk mitigation: Even if one central curator usually does a better job than the crowd, decentralization mitigates the risk that the central curator gets occasionally affected by perverse incentives.
Advantages vs. Central Platforms
- Sock-puppet resistance: On central platforms curators can open a large number of accounts, thereby diluting existing curators. In TCRs token supply can be designed such that it’s fixed, thus ensuring that governance rights are a scarce resource.
- Censorship resistance: TCRs are always open to all applicants, curators, and consumers.
- Transparency: Votes can be designed to be fully transparent. The audit trail of every vote is recorded on an open blockchain.
Advantages vs. Free Curation
- Skin in the game: Curators are also token holders. In a well-designed TCR where the quality of curation is linked to value accrual to the token, token holders are incentivized to curate a useful list.
- Anti-spam: A nonzero ticket price means that not everyone can apply.
- Marketing and R&D budget: In many cases TCR developers are also curators. Income that they earn by participating in the system can be deployed in marketing and R&D that could further increase the value of the TCR.
It’s noteworthy, conversely, that the value proposition of TCRs is reduced under the following circumstances:
- There already exists a reputed central curator.
- The central platform can be trusted.
- The three groups of participants don’t want economic transfer, e.g., applicants don’t want to pay to be included on a list.
In the real world, these circumstances are usually present. As such, we’d venture to say that TCRs are unnecessary in the majority of curation use cases.
Known Problems and Practical Solutions
TCRs may be a useful tool for certain applications, but the economic design and parameters will always vary based on the specific business use case. Consequently, it’s nonsensical to be bullish or bearish about this class of tokens as a whole. Each proposed TCR will need to be evaluated in the context of the problem it aims to solve.
However, we have identified a few common problems that likely affect almost every application of TCR we’ve seen so far. Here we describe these problems and how we think about solving them.
TCRs are largely social signaling systems. If no one is on the registry, then no one wants to apply to the registry. If no one applies, then there no one is on the registry.
Bootstrapping the registry therefore requires a ton of business development work. Every early applicant wants to hear that a few other highly qualified have at least shown some verbal commitment. A TCR developer’s job is thus to begin by speaking with a few potential applicants who are most likely to support the project and presenting to them a set of compelling reasons for joining the registry.
In addition, we can build an incentive structure for early applicants, and give them an additional stake in the system itself. One way to do this is to reserve, say, 20% of the total token supply for successful early applicants, and allocate 10% to the first 10 successful applicants, 5% to the next 10, 2.5% to the next 10, and so on. This could give early applicants enough financial rewards to cover their applicant fee, and governance rights over future applicants.
Lack of Price Discovery and Stability
In the early days of a TCR, the native token may not have a liquid market. Applicants may not be able to acquire tokens to pay for application. Applicants and curators face uncertainties with respect to how much value is at actually stake. These are mental hurdles for early participants which aggravates the chicken-egg problem.
A potential tweak would be to let applicants pay in native cryptocurrencies or stablecoins that are interoperable with the TCR. Native cryptocurrencies like Ether are sufficiently liquid. Stablecoins like the Dai, if and when they work at scale, will reduce financial uncertainties.
Economic Attacks on and by Curators
There are a number of economic attack vectors on and by curators, which would harm the integrity of the registry. For instance, even if they already have skin in the game by owning native tokens, they can be bribed by applicants or other entities who don’t want certain applicants to be admitted. This will certainly happen if the benefit of admitting/rejecting an applicant outweighs the bribe. Another type of attack is that a few large token holders can collude to control 51% of the voting rights, thereby defeating the whole purpose of independent opinions.
What we have come to conclude is that it is highly difficult to fully mitigate these attacks if all we have is economic incentives. We have to bake identity into the system, such that curators put not just money, but also reputation, at stake. More specifically, token holders must go through a KYC process in order to claim voting rights; otherwise their tokens only give them rights to cashflow. This way we have a double incentive structure where economic incentives should keep curators honest most of the time and reputational incentives should alleviate risks of occasional economic attacks.
Low Voter Turnout
Consistently low voter turnouts make people question the legitimacy of the registry. There are a few ways to solve this problem.
First, initial distribution of tokens is critical. Tokens should specifically be distributed to entities who actually care about the particular curation task and have the expertise for it, rather than short-term speculators.
Second, we could use a vote delegation mechanism, whereby token holders are not rewarded if they don’t vote, but are rewarded if they delegate their voting rights to a third party who have the time and the expertise to vote.
Third, instead of letting votes take place on random days, use a regular schedule, e.g., every other Sunday. Let it become a habit, which will help reduce cognitive costs.
Subjective Schelling Point
In most TCR designs to date, voters get rewarded if they are in the majority bloc. The optimal strategy for them is then not to vote based on the qualification of the applicants, but vote on what they think other voters will vote on. This is a major problem if the truth being voted on is subjective, e.g., Michelin 3-star restaurants.
The fix is relatively simple — to reward all voters, instead of just those who are in the majority bloc. The obvious downside of this is that it could encourage the coin-flipping behavior, whereby curators cast an arbitrary vote and still collect reward. Again, the aforementioned identity system could help here.
Another way to overcome the inherent problem of lack of objective, verifiable truth, is to let a central entity curate the registry for the few months or years, and later decentralize themselves over time. This way they will set the culture for the community, and, in particular, the standard for what is considered an acceptable applicant. We’d argue that this is not only helpful, but necessary to building a valuable network. The Bitcoin community values things like decentralization, immutability, and deflation so much because its founder and early adopters did so.
In this essay, we first laid out reasons why TCRs may be worth considering as an alternative to their central and free counterparts. We then reviewed a set of major problems we have observed and anticipate to see, and described our solution proposals.
TCRs are still unproven at this point. Like many other concepts in this industry, they are overhyped relative to their current development.
At the same time, we do believe that like many other concepts in this industry, TCRs, and token engineering in general, are a worthy experiment, as they enable new incentive alignment models.
For this reason, we are excited to contribute our thoughts to this experiment. Digital cash was a multi-decade experiment before Bitcoin was born, and we would not be surprised that token engineering will also take many years of painful iterations, sudden breakthroughs, serendipitous connections between ideas, and collaboration between great minds, before something truly works at scale.