Re-Fungible Token (RFT)

What happens when a Bonded Token owns an NFT?

Billy Rennekamp
10 min readFeb 26, 2018

UPDATE: The Re-Fungible Token concept has been converted into an EIP draft that can be viewed here:

The point of this writing is to explore the potential outcomes of using an ERC20 Bonded Curve Token (as per Simon de la Rouviere) or Liquid/Smart Token (as per Bancor) as the owner address of an ERC721 Non Fungible Token (NFT). Instead of a specific individual owning a specific NFT, they could hold a quantity of fungible tokens that represent in some way the original NFT. I suggest this could have interesting applications in curation markets for Art, Intellectual Property and digital paywalls.

ERC20 & ERC721

To begin it might be good to go over the specifics of the tokens referenced in the Ethereum Improvement Proposal Issues ERC20 and ERC721. If you’re already familiar with these terms feel free to skip ahead. If not please take a look at this Token Lexicon I’ve written up as a supplement to this article.

Bonded Curves & Liquid Tokens

There’s a special type of ERC20 token that allows users to buy and sell directly to and from the token’s contract. Simon de la Rouviere calls them Bonded Curve Tokens and Bancor calls them Liquid Tokens. Both refer to a contract which will accept ETH or another EC20 token in exchange for minting new tokens of the contract’s denomination. The contract will hold that original ETH or ERC20 in reserve for when someone wants to sell the tokens back. While technically very similar the two terms represent different use cases.

Bonded Curve Tokens & Curation Markets

Simon’s version is named for the curve of the graph that represents the price per token (y-axis) bonded to the number of tokens in circulation (x-axis) by a predefined slope formula. The slope can be linear, exponential, logarithmic or completely arbitrary depending on the use case but most often the price increases with the number of tokens in circulation. His white paper can be found here.

These Bonded Tokens are mostly described for use within Curation Markets, where the price of a token can represent the popularity of a topic. One example is trend forecasting that could be conducted by buying and selling different tokens representing different trends. People predicting the success of a certain style might buy the tokens expecting more people to follow as the style gains notoriety. The price will subsequently rise with the newly minted tokens creating profit for early adopters. Similar to Prediction Markets, it extracts the “Wisdom of the Crowd” with financial incentives. Curation Markets have many different applications and are being developed by a variety of different companies.

Liquid Tokens & Market Makers

Bancor calls this same technology a Liquid Token, emphasizing the fact that the token is always available for buying or selling in any quantity. They identify the relationship between the buying denomination and the selling denomination as a reserve ratio. Ratios below 1/2 increase exponentially, ratios above 1/2 increase logarithmically and ratios at 1/2 increase linearly. The price of a purchase is calculated with the total supply of tokens, the amount of underlying reserve and the pre-determined reserve ratio. For more details please read the excellent technical article by Slava Balasanov below.

Bancor’s use case is for supporting micro-communities who benefit from their own token but don’t have the quantity of buy and sell orders to work effectively on a traditional exchange. By encouraging the use of their own Bancor Network Token (BNT) as the common buy-in denomination they can provide exchange between different Liquid Tokens. If Liquid Token ABC wanted Liquid Token XYZ, and both used BNT as the buy-in token, then ABC could exchange itself for BNT then turn around and use that BNT to buy XYZ.

Re-Fungible Token (RFT)

Finally we can get down to the thought experiment I’ve originally proposed: What would happen if an ERC721 NFT were owned by the address of an ERC20 Bonded Curve/Liquid Token? Essentially a non fungible asset would become fungible again. The minting and un-minting of these fungible tokens would be bonded to a predefined price curve (and hopefully follow the conventions suggested in ERC621). Ultimately it would be possible to prove partial ownership of an NFT in the same way that it’s possible to prove token ownership through the standard ERC20 interface. If I were a wallet, I could check the Digital Asset Registry (DAR) of an NFT to see who owned it. I could then check that owner address to see if it’s in fact an RFT contract. If so I could use the balanceOf function to check whether some address owns a quantity of that RFT token and verify partial ownership of the original NFT.

Depending on the scenario some signal could be gleaned by the price per token representing the asset. An RFT with a very high price must represent some desired quality in order to have so many users willing to buy it. These users may be rational actors who expect to extract a profit on further growth of the token or they may be altruistic supporters of the topic regardless of their economic outcome. The signal gained from the price per token might be different depending on the type of NFT it represents, as each use case demands.

Art in a Curation Market

It’s no secret that information asymmetry is rampant in the art market. This configuration largely benefits dealers and auction houses who are often able to leverage larger sums for the sales of a work by keeping prices, offers and buyers secret. Attaching NFT artworks to Bonded Curve Re-Fungible Tokens would create a transparent system in which the values of the artworks were constantly updated by the speculative Curation Market. While such a system might be ripe for click bait exploitation it could also provide a mechanism which benefits the artist. Investors with keen taste stand to generate an income through savvy acquisition. Artists meanwhile could similarly speculate or pre-mint their own tokens or leverage an inflation on the RFT which would provide a continued source of income. In the case of inflation there could be a spread on the buy and sell curves where the difference between them pays out continuously to the author of the work, or a simple flat fee on each exchange. Essentially the artist would make money every time the work was sold.

This scenario may create more questions than it answers. To what degree would the price of an artwork-specific token reflect the quality of that work? What would it mean conceptually if an artwork could only be owned by a collective? Would this result in more or less ethical consumption of art? More or less volatile prices?

As an aside, it might be interesting to see what would happen if curation markets in this context were nested. Each Bonded Token accepts some base denomination upon minting. This base denomination can be Ether, a stablecoin like Dai or any other ERC20 token including another Bonded Token. This system might signal multiple aspects of the artwork in our scenario. On one level might be a General Investment Token (GIT) bonded to ETH or a stable-coin. This GIT might be used to buy different Digital Asset Registry Tokens (DARTs) that could represent whole genres of NFTs; Imagine a CryptoKitty DART, a CryptoCelebrity DART and a CryptoPunk DART. The final level would be the Re-Fungible Token (RFT), purchased with DART and representing partial ownership in a single NFT (relevant to the topic of the bonded DART of course). Prices of the NFT bound RFTs should signal the success of individual artworks, prices of DARTs signal the success of whole genres of Digital Asset Registries and the price of GIT signals the success of the entire system. Simon outlines a similar scenario using nested Token Curated Registries in the article below.

Copyright Management

An RFT could be used to manage ownership of an NFT representing intellectual property (IP) while simultaneously providing signals about the popularity of that IP. Governance could be handled in various forms — with or without the use of the RFT — for deciding things like who can use the IP and how much it would cost. Similar to the artwork scenario, a continuously mintable token would allow the creator to benefit by self-speculating, self-assigning a pre-minted quantity of tokens, or through an inflation mechanism that pays directly to the author over the lifetime of the token.

Payments for the usage of intellectual property could be paid in the denomination of the token, further raising the value. These payments might go directly to the token founder (and apparent IP author), directly to the collateral pool (raising the price per token without increasing the number in circulation) or to an inaccessible address like 0x0 (raising the floor of minimum price per token by preventing those tokens from ever being sold again).

Digital/Physical Paywall

Sol LeWitt’s wall pieces are texts with instructions for their creation.

In the first example the ownership of an artwork was represented by an NFT and in the second it was intellectual property. Both situations use the token as a stand in for content that exists elsewhere digitally, physically or conceptually. Both leverage a potential inflation mechanism to pay the author when speculators make money off of the token — essentially monetizing attention. As mentioned earlier, a Re-Fungible Token would provide an API for verifying partial ownership of a non fungible token. If this NFT were either an artwork or a form of intellectual property, the verification could be used in a paywall.

Preventing pirating of copyrighted material is a cat and mouse game that shouldn’t be pursued. Time is better spent offering easy access to high fidelity content at a subscription price. In the case of music this has already occurred with promising success but it might be improved by use of an RFT. Consider the album that drops and is only accessible to download and stream through a service that can verify you are a token holder representing that album. You’d buy tokens because you want easy access and you know that you can sell the tokens back when you’re tired of listening to it. Alternatively ownership of these tokens may give access or discounts to physical events. Either way the price of the tokens would rise with the popularity of the music. Inflation could ensure that the musician makes money on all of these transactions. The musician would also be smart to invest at the launch of the token; being the first buyer ensures that the price will never drop below what was paid. In this scenario, fans stand to earn alongside the creator and will do more to support their mutual success. A similar model has been proposed by (again) Simon de la Rouvier in which the tokens would be used for buying the actual album instead of as paywall.

This may be the unfortunate beginning of the “steem-ification” of music. I worked at my college radio station and had to field phone calls from promoters in order to get free copies of new albums. Along with the stuff we wanted came the garbage they were paid to sell us on. It was a really toxic experience and turned me off of new music for a long time. I realize essentially paying fans to promote music might end up similarly sour ☠️

What’s next?

All of these scenarios will need a lot of experimentation as there are a lot of unanswered questions. What parameters would be needed to create an efficient system? How many tokens should be required for paywalls? Should that amount change over time? How should the price curve be set? What level of inflation provides adequate payment for token founders while providing attractive investment opportunities? Are we setting up a scenario for vicious pump and dumps? What is the legal status of these tokens?

💡 I hope this article has stirred your interest in this topic. If you have other ideas for a Re-Fungible Token please comment about them. What would happen if the NFT were the deed to 🏠 a house? 🚗 a car? ⛵️ a boat? ✈️ a plane? 🏝 a piece of land? 🐶 a puppy?

For other writing on similar subjects I’ll try to keep this list updated:

Thank you to Matt Condon, Simon de la Rouviere and Slava Balasanov for feedback and edits.