Financial market data network Pyth has unveiled an incentive scheme for its providers – typically high-speed traders and native crypto… Read More »
In a white paper published this week, Pyth makes the regular DeFi industry argument about market data being available only to a “limited set of institutions and users”, while reinforcing its intention to deliver its data to the general public. The network aggregates market maker and exchange price data on-chain and makes it available for use by either on- or off-chain applications.
Under the scheme, end-users of Pyth data can voluntarily pay data fees to protect against potential oracle failures, namely wrong data being published, as was the case on the network in September 2021 when it erroneously reported a 90% drop in the price of bitcoin. So-called Delegators, effectively owners of crypto tokens who can use those coins to back data providers, are able to choose which price feed and specific publisher to back, in return they will earn data fees, or, if wrong data is published by the firm they back, they lose their stake (assuming it is a fault at the publisher level). A Pyth token is being rolled out gradually for use by delegators and other stakers – and participants can have multiple roles within the network, specifically, publishers or consumers may delegate additional tokens to increase the fees they earn.
The Full FX View
The idea of rewards for providing data is a radical one when looked at from the confines of traditional markets, for years now certain institutions in FX markets especially have wondered aloud why they spend so much on data, a good deal of which comes from their own organisation.
Hence the Pyth idea is an interesting one, but it also makes certain assumptions. The network has gathered significant support from high speed traders and native crypto market makers, but it is by no means universal. This means the incentive to voluntarily pay for market data is undermined to a degree – after all, who willingly pays any fees?
Much has been made of one or two instances towards the end of 2021 when data sources (including Pyth) recorded some serious outliers in the prices of crypto assets, but considering across the other 363 days of the year there seemed to be no problem at all, is the economic argument convincing enough?
Pyth itself says in the white paper that consumers are incentivised for two reasons – data fees enable applications to reduce the risk of using Pyth price feeds as they would receive a payout in case of failure. Second, paying data fees attracts more publishers to the product, which improves the robustness of the price feed. Is this going to be enough? More pertinently, does Pyth run the risk – which its protocol seeks to neutralise it should be said – of attracting too many data providers of lower quality? Even in such a huge market as foreign exchange, the bulk of the data emanates from less than a dozen participants – just look at how the FX committee surveys show that 90% of volume is seen by 12 players.
Equally, Pyth is trying to promote itself as a provider of data across all financial markets, not just crypto – however it will be challenged to get end users to pay for market data in TradFi when a suitable feed (for their uses anyway) is available from the trading venue or venues they already use.
The idea of providing rewards for data provision has circulated for some time in crypto markets, however to date, in tradition finance the hold data providers have on their markets has remained unshaken. Putting aside the irony of how a DeFi proponent is seeking to centralise data, it will be interesting to see how this project goes.
There are still risks that a small group of players end up dominating the network, but should we be worried about that? Generally speaking in crypto markets, any market manipulation seems to come from sources other than these independent market makers and venues, so if there is a data source from this group it should be embraced – although voluntary payment may be a little too optimistic in the long run.
The goal of the scheme, Pyth says is to make its network self-sustaining and decentralised.
The protocol is designed to reward publishers of data in proportion to the quantity of new pricing information they share on the network. It consists of four on-chain core mechanisms, price aggregation, data staking, reward distribution and governance.
The aggregation feed combines the reported prices and confidence intervals into a single feed, designed, Pyth says, to ensure that a small group of publishers cannot “significantly influence” the price.
The delegators’ tokens also determine the level of influence that each publisher has on the aggregated feed – this raises the prospect, of course, of one or more firms acquiring high influence, which could, feasibly, undermine the first aggregation logic. Of the fees collected from consumers, 80% is distributed to delegators, which 20% goes into multiple reward pools for each product to be distributed among publishers.
The protocol governance will use a coin-voting system, that will decide what tokens may be used for data fees, the products listed on the network, the share of data fees allocated to publishers and the number of Pyth tokens that need to be staked to enable a claim for erroneous data.
Effectively, if a publisher posts incorrect data that impacts the aggregated data feed, their stake will be “slashed” and the tokens given to those who make the claim. The latter, as hinted above, will have to post a bond, to “prevent spam” Pyth states in the paper. To check whether a data feed was wrong, Pyth says it will collect off-chain information from “impartial judges” and then feed that information into an algorithm that determines the outcome of the claim. Token holders then have to ratify the algo’s decision.