Information Asymmetries, Adverse Selection, and Token Markets
In 1970, renowned economist George Akerlof famously introduced the world to the “Market of Lemons” in which he discussed how asymmetric information among agents in a market may lead to a degrading quality of products — subsequently market failure. When applied to the current situation in the token market and development of DAPPs, one can infer that blockchain technology is facing issues of tremendous importance for future (mass) adoption.
Adverse Selection & Token Markets
In his paper, Akerlof takes the example of the used car market, which we adapted to the token market:
Assume a potential token purchaser (Dₓ) and three potential projects to offer a cryptocurrency (Sᵦ , Sᵧ, and Sᵨ). It is easily to be inferred that the projects’ founders have by far more information on the potential/quality of their cryptocurrencies to be sold (Cᵦ, Cᵧ, and Cᵨ) than the buyer does — especially as the projects’ founders know whether or not they actually plan to conduct a scam rather than develop an innovative technology.
Let us further assume that the projects seem identical on the outside, however, seller Sᵨ’s project is actually a scam. Moreover, the projects’ founders know the honest (fair) value¹ of their respective coins, which are V(Cᵦ)=1, V(Cᵧ)=0.5, and V(Cᵨ)=0. Sᵦ, Sᵧ, and Sᵨ now set prices, whereas Sᵦ and Sᵧ are honest and sets the price at P(Cᵦ)=V(Cᵦ)=1, P(Cᵧ)=V(Cᵧ)=0.5 respectively. Sᵨ, who knows that his/her project is not visibly different, though a scam, is trying to maximize profits in anticipation of Dₓ’s decision making process. Assuming Dₓ knows the underlying probability distribution, the agent will set a willingness to pay based on the expected value of participating in the market in the first round. In our case this would result in E₀(V(C))=0.5. Accordingly, Sᵨ will — in order to maximize profits — set the price to P₀(Cᵨ)=0.5=E₀(V(C)).
The result of this market situation would be that Dₓ will transact with either the malicious seller Sᵨ or the low potential seller Sᵧ, however no contribution will be made to Sᵦ. Further applying this to n agents, both on the demand side and supply side, who are faced with similar options, it would lead to any honest, high potential agent — such as Sᵦ — being driven out of the market and any agent D facing losses by being scammed or ending up with a low quality coin. Over multiple rounds, this would then lead to the eventual market breakdown, as only low potential projects and scams remain in the market for the subsequent round. This dynamic is further driving down the willingness to pay, from E₁(V(C))=0.25 in round two to E₂(V(C))=0 in round three, after which no exchange is facilitated anymore since any remaining project in the market is a scam.
Web 2.0 Solutions to Adverse Selection
This problem is no novelty, yet, posed tremendous difficulties in the early days of the internet, as users were not able to assess whether any seller or service provider actually existed or provided its services as promised. Many times users were scammed with non-existing product offerings on eBay and similar websites. However, eBay was amongst the first to tackle this issue by sticking to its fundamental belief in network effects and peer-to-peer platforms: the introduction of a rating systems, in which purchasers were able to review and assess retailers. eBay as the central platform provider would subsequently aggregate and verify those ratings, then post them to the retailers offering and profile.
With this mechanism, eBay was able to resolve some of the information asymmetries in Web 2.0 — a novelty that has become a standard nowadays, with Amazon Ratings, Google Reviews, or Twitter’s Verified Profiles.
Web 3.0 Problems with Adverse Selection
With the emergence of blockchains and DAPPs, we were promised trustless exchange and immutable databases, however, are in fact facing the exact same problems early internet applications and businesses faced — information asymmetries.
One may raise the question of:
- Why an OpenBazaar is not posting fault-tolerant ratings on sellers?
- Why ETHLend is currently not evaluating the credit score of a new borrower, but requests them to pose collateral?
- Why Prediction Markets, such as Augur or Gnosis, have difficulties resolving their respective markets at termination of an event?
- etc.
In a widely discussed blogpost, Terence Eden perfectly trolls Verisart — a blockchain application aiming at protecting copyrights of artists — and thereby unveils one of blockchain’s most fundamental problems. More specifically, he was able to claim copyrights to the Mona Lisa and stored it immutably on the Verisart blockchain. Hence, he made clear: blockchains are great at keeping an immutable, byzantine fault-tolerant record of transactions, though blockchains are less optimal at evaluating information outside of their own system, as no central entity exists to do so (as opposed to Web 2.0).
Rlay: The Web 3.0 Solution to Adverse Selection
Without a central entity evaluating individual participants claims concerning all information outside the closed blockchain’s system, blockchains are prone to being immutable databases of misguided transactions. Anyone can claim to be Leonardo Da Vinci or claim to have a perfect credit score.
With Rlay, we are providing the first truly decentralized, efficient, and attack resistant solution for adverse selection. We developed Rlay to prevent smart contracts from generating misguided transactions based on wrong real-world information and enable apps and DAPPs to use Rlay as the first decentralized information source with the same efficiency and attack resistance guarantees as other blockchains, to prevent token and other markets from adverse selection and potential market failure.
We will be releasing more information in the coming days and weeks, including our testnet on OST & Ethereum. If you are interested in discussing use cases for Rlay, feel free to reach out to us and join our community, via one of the links below!
Thanks,
Sam
Disclaimer
This piece constitutes an independent article and does neither pose investment advice nor any claim that any cryptographic tokens and coins in itself pose an investment opportunity. This article reflects the author’s individual opinion and does not necessarily represent the opinion of the Rlay project, Project T Ltd., or any other affiliated party.
1) This is a general assumption of Akerlof’s model on adverse selection and does not imply that any cryptographic token or coin is having “intrinsic value”.
