IIP-091: Launch the NFT Blue-Chip Index (JPG)

IIP: 91
Title: The NFT Blue-Chip Index (JPG)
Status: Proposed
Authors: @JosephKnecht (Article21), @oneski22 (Index Coop)
Discussions-to: [Discussion] NFT Blue-Chip Index (JPG)
Created: 09 Sep 2021

Simple Summary

The NFT Blue-Chip Index (JPG) is a market cap-weighted portfolio of blue-chip NFT collections. The index provides diversified, liquid exposure to the top NFTs through a single liquid token. According to Index Coop, an NFT Index token is its most frequently requested product and this token can address that demand.

Abstract

JPG is a liquid, market cap-weighted index of NFT collections. The index is composed of NFT liquidity vault tokens such as NFTX, NFT20, and Fractional. The token is based on the NFTX Vault Index token (NFTVI) which fair-launched on August 28, 2021.

The JPG token can be implemented using the Simple SET Protocol provided that only the NFTX and NFT20 protocols are used. If the Fractional protocol is also used, the manager will need to be able to claim the ETH proceeds from auctions. Alternatively, the provision and redemption of NFTs to Fractional could be managed by a separate contract with the fractional shares then supplied to the index contract.

In order to reduce price impact, it would be desirable for the index contract to be able to selectively redeem vault token components for NFTs and conversely to mint NFTs from vault tokens. For example, the index contains PUNK tokens. One PUNK token could be redeemed for a Punk NFT on NFTX and then sold on OpenSea for ETH. Similarly, ETH could be used to purchase a Punk NFT which could then be used to mint PUNK tokens. If these mint/redeem functions are too difficult to implement then we would need to raise additional vault token liquidity.

It is envisioned that supplying vault inventory and creating new collections would be managed by a separate address from the index token address. The two addresses would not need to interface with the exception of providing the vault tokens to the index contract. We do not foresee using synthetic derivatives such as uPunks from YAM synths.

Motivation

Rationale

Non-Fungible Tokens (NFTs) are a revolutionary cultural, artistic, and financial phenomenon. However, despite the massive size and growth of the NFT market there is no NFT index token.

For institutional investors, benefits of an NFT index token include access to a highly illiquid and fragmented market, a single instrument for long/short positions, enhanced diversification, faster order fulfillment, reduced transaction fees, and automatic rebalancing. Advantages for retail investors, in addition to the above, include easier onboarding, mitigating unit bias, and lower capital entry requirements.

Creating an NFT index token is highly challenging however due to the illiquidity, volatility, fragmentation, and price opacity of the NFT market.

Specification

Overview

JPG will be a market cap-weighted index of NFT liquidity vault tokens. Additional functions may be needed to mint and redeem NFT vault tokens and buy and sell NFTs on the secondary market. It is an engineering decision whether this is done from the index contract address or a separate contract.

Differentiation

As the first NFT index token on the Ethereum blockchain, JPG is a highly differentiated product. The JPG token is unique in providing liquid access to the NFT market at scale. The product does not overlap with any existing Index Coop products and could be cross-sold with MVI.

Other nominal NFT index tokens such as NFTI and NFTP hold NFT protocol tokens (ie, AXS, ENJ, SAND, SUPER, etc) and not NFTs themselves. One potential, future overlapping token is the PFP (profile pic) index planned by PieDAO. However, this product will only cover PFP NFTs and the proposed holdings include many highly-illiquid microcaps. Zilliqa recently launched an NFT index token $NFTDEX; however its components are restricted to NFTs on the ZIlliqa blockchain. Strudel Finance has an NFT index consisting of a Balancer pool of 25% each of wBTC and the NFTX vaults PUNK, BAYC, and COOL. However, this pool is diluted by wBTC, is not diversified, and has a constant allocation.

JPG is based on the NFTX Vault Index (NFTVI) which fair-launched on August 28, 2021. As of September 18, the Net Asset Value (NAV) of NFTVI is up 69% since inception. See Zerion for NAV performance and index composition.

Article21 is the author of the current proposal and also the creator of NFTVI. If this Index Coop proposal is successful, Article21 will retire NFTVI and be fully dedicated to the success of JPG.

Example composition

Performance

See the historical NAV performance of a similar index token NFTVI.

Size of Opportunity

The market cap of the top 20 NFT collections is $9.4B and growing extremely rapidly. Assuming 5% of the top 20 market cap can be made liquid, accessing 25% of that market would represent a $117M opportunity for Index Coop. Additionally, Index Coop could charge a premium management fee given the uniqueness of and exceptionally high demand for an NFT index token.

The size of the opportunity will be determined by how much resource and risk IC wants to invest in liquefying top NFT collections. Currently, vaulted collections typically only represent <0.2-1% of the collection and many top cap collections are not vaulted at all.

We predict AUM of $50M, $200M, and $500M at 6, 12, and 24 months respectively. The AUM will depend heavily on how much resource IC invests in liquefying top NFT collections. The market cap of the top 20 collections is $9.4B and growing rapidly.

Market & Customer Research

Target Customers

Institutional investors seeking straightforward, liquid access to the NFT market at scale.
Retail investors seeking to enter the NFT market but lack the resource, time, or risk appetite to invest in individual NFTs.

User stories

Private equity shop Alpha has a long NFT thesis based on emerging cultural trends. They are crypto savvy and have crypto holdings, but they are not interested in purchasing individual NFTs. They purchase the NFT index token OTC for broad diversification, fast fulfillment, and liquidity.

Family Office Beta has a mandate to invest 25% of their client’s portfolio in art, 50% of which can be alternative assets. They have taken notice of high-profile NFT auctions and recognize the nascent asset class, but are uncertain how to access the market as they have no familiarity with crypto. They purchase the NFT index token because of the blue-chip collections they recognize from auction, the diversification and liquidity, and the trust and support provided by the IC institutional business team.

Retail investor John is excited about NFTs as a new cultural phenomenon and asset class but doesn’t have the wherewithal to invest in individual NFTs. In particular, he’s worried about the lack of diversification and exit liquidity. He has DPI and MVI and likes the idea of applying indexing to NFTs. He buys the index token because of the simplicity, diversification, automated rebalancing, lack of correlation with other crypto, easy tradability, and the strength of the other Index Coop products.

NFT enthusiast Jane is an avid NFT collector. She has several grails and also picks up emerging collections. She’s skeptical of the financialization of NFTs and the blue-chip designation. However, she’s interested in JPG because she recognizes that it’s a first-of-a-kind project, and the team is positively engaged with the community.

Methodology

Initial Composition & Token Inclusion Criteria

The JPG index token is composed of NFT collections represented by NFT liquidity vault tokens. The NFT collections are selected based on being top floor cap and available through a liquidity vault. The floor cap (defined as the OpenSea floor price times the total collection supply) is used instead of the market cap since floor cap data are more readily available, tend to be more reflective of the liquidity vault contents, and the two are highly correlated regardless.

The NFTX and NFT20 protocols were chosen for the liquidity vaults because they are well-loaded, fully-collateralized, mainly liquid, fractional, and DAO-governed. NFTX covers most of the largest cap NFT collections. Note that one can represent a collection with multiple vaults in order to increase diversification and mitigate risk, or to compose a vault group with specific properties.

Selection criteria for collection

  • Floor cap > $50M
  • Art only. No land, audio, or domain names. (We will use an on-chain oracle to determine what is art … just kidding)

Selection criteria for liquidity protocol

  • Liquidity vaults representing the above collections. Current liquidity vaults include NFTX and NFT20
  • Fractional NFTs and synthetics TBD
  • No individual NFTs

Selection criteria for components

  • NFTX or NFT20 vault token
  • Liquidity
  • Inventory
  • Market cap
  • Complementarity needed to compose a net position for a collection
  • Available from SetProtocol (i.e., ERC-20 standard, listed in the CoinGecko Token List, sufficient liquidity on Uniswap or Sushiswap).

Weightings

The index components are weighted by the square-root of their collection floor caps. In the general case where a collection can be composed of multiple grouped components, the weighting for the ith component of the jth collection is:

where fj is the floor cap of the jth collection, ωij is an additional weighting of the ith component of the jth collection and n=½. The additional weightings ω can be used to combine components with, for example, different quality levels or attributes (Figure 1). Note that the market caps of the vaults themselves do not play a factor in the weights.

Square root-weighting is used as opposed to linear weighting due to the very heavy concentration of the floor caps. For example, linear weighting would result in >70% of the index being in the top 2 collections, ie, CryptoPunks and BAYC.

Figure 1. Illustration of how the index is composed of collections, eg, CryptoPunks, CyberKongz, and so forth. The collections are in turn composed of liquidity vault tokens (from NFTX and NFT20). The vault tokens are fractional claims to individual NFTs. The components are weighted wij by their collection floor caps fj and the component weightings ωij.

On-Chain liquidity analysis of underlying tokens

The current liquidity on most NFTX vaults is extremely low to the point of being effectively zero for our purposes. NFTX will be remedying this by allocating 25% of all future revenue to liquidity. The timeline for rolling this out has not been set. One factor in the low liquidity is that a lot of the trading volume is done by minting and redeeming NFTs out of the vaults as opposed to swapping the vault tokens, eg, see the PUNK Dune Analytics.

It’s also worth noting that the price impact on selling component tokens would effectively be capped at ~5% since this is the fee to mint NFTs from the vault tokens. The effective price impact is not 5% exactly because of the margin between the token value and the NFT floor price on the secondary market. Similarly, the effective cap on buying component tokens would be the margin between the vault token price and the secondary market price.

Here are the current empirical trade sizes that would give 1% price impact per component:

Based on the allocations above, a 1.3 ETH trade (!) on the index is enough to induce a 1% price impact in $TOADZ, followed by a 1.5 ETH trade for WOW.

Maintenance

The index will be rebalanced monthly to reflect the target allocation and to consider new collections and liquidity protocols. Urgent rotations will also be conducted if the inventory is depleted in a liquidity vault or a fractional holding is successfully auctioned.

The non-linear weighting increases the tracking error so we propose a relatively high Maximum Absolute Percentage Error of <5%. Advantages of the higher threshold for rebalancing also include reduced price impact and transaction fees.

Costs

Cost to customer

To be determined after DG1.

We anticipate JPG having a 3-5% management fee given the high demand and uniqueness of the project, as well as the effort, capital, curation, and risk required to manage the vault liquidity and inventory.

It is worthwhile considering whether we can pass on the 5% vault token minting fee to minters of the index token. Also, we can consider allowing users to redeem the index tokens down to the underlying NFTs for an additional fee. This will help communicate that users own a claim on the actual underlying NFTs.

Rebalance frequency

Monthly

Manual Rebalance magnitude

We anticipate trading 3-4 positions at each monthly rebalance.

Fee split

To be determined after DG1.

Meta / intrinsic productivity

Not applicable

Liquidity

The trading pair for JPG will be JPG:ETH on UniSwap V2. The liquidity for the vault tokens will be automated, concentrated liquidity such as G-UNI (ie, Gelato Uniswap v3) or Visor-Sushiswap-Trident. The advantages of the improved performance of the concentrated liquidity and the trading fees will be offset by the much greater impermanent loss. The tradeoff will depend on the effect of price impact drag on TVL uptake, trading volume and the details of the G-UNI automation. Consequently, it may be useful to conduct experiments before launch or do a supply-capped launch.

It is not necessary for our vault LP platform to match that of the existing vault tokens since the 0x exchange used by SetProtocol will split transactions across multiple exchanges to get the lowest overall price impact. Gelato-Uniswap v3 LPs are ERC-20 and so can be pooled.

Concentrated liquidity is needed because the amount of non-concentrated liquidity needed would be exorbitant. Consider an existing liquidity pool consisting of trading pairs x and y. From the Uniswap constant product formula (x+Δx)(y-Δy)=k, one can derive that the liquidity x needed to achieve a price impact P for a trade-size of Δx is x=Δx (P+1)/P.

We consider a scenario of JPG having 3,000 ETH TVL and needing to completely exit a 3% (=90 ETH) holding. In order to achieve a 1% price impact, from the above formula, the liquidity needed would be 9,090 ETH plus 9,090 ETH equivalent of the other trading pair (minus the existing liquidity which we take to be negligible) which is clearly excessive. For 5% price impact, we would still need 1,890 ETH + 1,890 ETH equivalent in liquidity which is still excessive. These liquidity requirements are exorbitant and so we anticipate needing automated, concentrated liquidity for the vault tokens.

For very large orders or mints, buy-side liquidity is anticipated to be available through floor-sweeping tools such as genie.xyz and NFTX. For NFTX, the effective price impact will be capped at ~5% for the minting fee.

Author Background and Commitment

@JosephKnecht is a methodologist at Index Coop and Founder of Article21

Article21 is committed to developing the next generation of crypto index strategies. A21 developed and released NFTVI, the NFTX Vault Index token

Marketing support / distribution / partnerships

The methodologist will promote the token through the NFTVI Twitter account, which can be rebranded and repurposed to promote JPG. The methodologist will also be available for AMAs and interviews.

References

Helpful NFT floor/market cap leaderboards:

NFTs as an asset class: Alternative Investments in the Fintech Era: The Risk and Return of Non-fungible Token (NFT) by De-Rong Kong, Tse-Chun Lin :: SSRN

Revision history

29 Sep 2021 - Expanded based on new product IIP Template in preparation for DG1 vote.

Copyright

Copyright and related rights waived via CC0.

26 Likes

A warm welcome to you @JosephKnecht and thank you for dropping this very interesting proposal.

I have been looking for something like this for a while, thought it might be impossible (due to liquidity and market structure issues) and am excited to see your thinking.

A few comments and questions:

  1. Why square root not cubed root for this methodology? I note that NFTVI currently uses cubed root, arguing the below, but I’m sure you’ve thought about this a lot.

*Weightings *
The index components are weighted by the cube-root of their collection market-caps. Cube-root weighting is used – instead of linear or square-root weighting – due to the very heavy concentration of the market caps. For example, even square-root weighting would result in >50% of the index being in the top 2 tokens.) uses cubed root.

  1. What do you think the management fee would be approximately to make sense for IC, methodologist and clients? I was thinking maybe 2% or 2.5% maximum would be reasonable from the client side. Crudely about 2x the holding cost of DPI felt right.

  2. What brings you to want to partner with the Coop?
    I hope we can do something here, but would transparently like to understand the alignment of incentives given this is quite a different index to launch and support (I think) compared to our other sector/themed index products.

  3. Does this index product require much new engineering work versus our other sector/thematic index products such as DPI, MVI, DATA? Tagging @dylan @ncitron into this too. The NFT market structure mainly makes me ask this question.

  4. (If you’re able to opine) What competitor risks do you see emerging in the market that we should be aware of regarding someone (finally) building an elegant and robust index for NFTs?

Thanks for your time and responses. Again, I’m so glad to see this post.

I’m very interested to learn @overanalyser @puniaviision 's thoughts on this proposal too.

3 Likes
  1. Also @JosephKnecht given liquidity is such an issue, how can Index Coop:
    a) help liquify this market efficiently?
    b) think of potential TVL for such a product over three month to two year horizons?
1 Like

Thanks. It’s wonderful to be here.

The very low liquidity and market fragmentation are definitely the main challenges. “Low liquidity” in this context refers both to the low liquidity and low inventory in existing vaults as well as the complete lack of vaults for some collections, e.g., VeeFriends, Loot. We’re working closely with NFTX on creative solutions to improve the liquidity in the existing NFTX vaults and there are some potential solutions to create vaults for collections that are not currently covered.

Powered-weighting is used generally to flatten the composition weights and reduce concentration risk. Lower power (ie, 1/3 instead of 1/2) has the advantage of flatter weights at the expense of more rebalancing and hence higher price impact. In the case of $JPG, the much larger market cap envisaged compared to $NFTVI puts a higher priority on liquidity and hence higher weighting on the larger caps with deeper liquidity. The net effect is less price impact on rebalancing. Also, from a marketing perspective, square root-weighting is potentially more palatable and consistent since it’s used in MVI.

We haven’t discussed the management fee yet so I’d prefer not to venture a range here. Generally speaking however I think we can charge at the higher end since its a highly differentiated product and there should be a significant liquidity premium. I’m a newcomer to IC but I also think you should be charging a higher management fee across all of your products. I’m not aware of any willingness-to-pay market research that informed the current fee levels.

We were delighted when IC BD approached us to port our methodology to IC. As you’re probably aware, an NFT index has been IC’s most requested product.

Generally, I see our skillsets and goals being very complementary. We have competence in investment strategy and portfolio construction and IC has deep capabilities in distribution, engineering, marketing, partnering, governance, and much much more. It’s far more time and capital-efficient for us to partner with IC than to build these capabilities internally. At the end of the day, we’re wonks who like to build products and we’re delighted to partner with IC to bring those products to market.

We’re also highly encouraged by @BigSky7’s desire to build a long-term partnership. We have other ideas for future products that we’ve shared with @Financial-Freedom, @oneski22 and the BD team. We hope to pursue those as soon as the NFT Index gets sufficient momentum.

Yes and no. The index itself is a ‘simple index’ in IC’s terminology but some additional work may be required outside of the index to improve the liquidity and coverage of the vaults. We’re working closely with @overanalyser and the NFTX engineers to assess the technical options. It’ll be a commercial/technical resource trade-off on how much price impact we’re willing to accept versus investing the partnering and engineering resource to improve the liquidity of the underlying vaults.

Our biggest competitor is time time time. The NFT market is growing so unbelievably fast that every week we delay we lose opportunity. In terms of ostensibly overlapping products, nominal NFT index tokens such as NFTI and NFTP hold NFT protocol tokens (ie, AXS, ENJ, SAND, SUPER, etc) but not NFTs themselves. The PFP (profile pic) index planned by PieDAO will only cover PFPs and the proposed holdings include many highly-illiquid microcaps. Zilliqa recently launched an NFT index token $NFTDEX; however its components are restricted to NFTs on the ZIlliqa blockchain. Strudel Finance has an NFT index consisting of a Balancer pool of 25% each of wBTC and the NFTX vaults PUNK, BAYC, and COOL. However, this pool is diluted by wBTC, is not diversified, and has a constant allocation. Chainlink is working on a punk floor pricing oracle for JPEG’d that someone could derivatize.

Whatever happens, an NFT index is a huge opportunity and I want IC to be the one to seize it.

It’s my pleasure. Thanks for the thoughtful questions. We’re looking forward to working together.

7 Likes

There are two aspects to increasing the liquidity: (1) increasing the liquidity and inventory of existing vaults (eg, NFTX CryptoPunks) and (2) creating vaults for collections that are currently not vaulted (eg, Fidenza or Loot) and hence not included in the index at all.

Liquidity (1) can be addressed with, for example, concentrated LP like Sushi-Trident or Gelato-Univ3 or floor sweeping NFTs outside the vault for large buy orders. Also, IC could incentivize LP creation for the vaults as was done very successfully recently for NFTX’s PURR collection.

Liquidity (2) requires more of a strategic and financial commitment from IC but would entail IC being the cornerstone or significant co-investor in a new vault collection, eg, Fidenza or zombie punks. Obviously, that would have a lot of implications so I mention it here only as food for thought.

Let me defer that question to IC BD because it depends heavily on how much resource we invest in (a). The current price impact calculations are based on an initial $10M TVL but I think we can aim higher.

2 Likes

This makes no sense as NFTX pool liquidity is unlocked. Meaning that in each of these pools listed (almost all proposed are NFTX), the NFTs can be pulled from them. This will only work using locked custom pools on a platform like Fractional. Otherwise, I’d suggest only protocol tokens.

2 Likes

You make a very important point @Beanie. You’re highlighting the very real risk that the inventory in an NFTX pool can be depleted, rendering the pool useless. To address this, we will have a minimum inventory requirement. This is why, for example, the current index doesn’t include NFTX BAKC since there are only 6 items in the vault. In another example, in the current index, we had to leave NFTX KONGZ and switch to NFT20 KONGZ20 because the NFTX vault was depleted. The importance of inventory was one of our key learnings.

NFTX is also working on an aggregator where if a buy order exceeds the vault inventory, then the difference is purchased from secondary markets. Having multiple vaults for a given collection (incl Fractional) may also help mitigate the risk of a run on inventory. Thanks again for highlighting the importance of considering stock as well as liquidity!

2 Likes

Just use Fractional. Problem solved. Why use a protocol that isn’t built for this?

1 Like

Using only Fractional is a very interesting idea. The main reason we’re considering other protocols at this stage is that there are some high cap collections that don’t have multi-item vaults in in Fractional. For example, Glyph has 0 vaults, Squiggle 0, Meebits max 1-item vault, Cool Cats max 5-item vault, etc. The transaction fees get very high if we have lots of vaults per collection. One interesting approach would be for IC to purchase these (or a set of floors) and re-deposit them as a multi-item vault. We’ll definitely discuss this with IC leadership as part of the project. I imagine with Fractional we also have the risk of a successful auction and we’d then have to redeploy the ETH proceeds. I hope that all sounds sensible.

3 Likes

Thanks for the responses @JosephKnecht. Most appreciated.

Quick thought re point 2) on liquidity: I wonder if some of our market making partners would be better placed to, well, make markets.

Cc @fallow8

Definitely. There are a few groups (eg, genie.xyz) developing multi-market floor-sweeping tools that should make that a lot easier.

1 Like

Just chiming in to signal my support for this proposal. Clear product differentiation and allows us to innovate on-chain with the Set team. This also allows our DAO which is overseas the biggest indices in crypto to support a product that covers the most popular segment in crypto, NFTs.

7 Likes

The Community Call has been scheduled for Wednesday Sept 29 (17:00-18:00 UTC). Dial-in Details. We hope to see you there, JK

2 Likes

@mel.eth Thanks again for today’s community call. I’d like to request an IIP number please. Cheers. cc: @oneski22

1 Like

Maybe dumb question but why not $JPEG instead of $JPG?

1 Like

Folks,

I’m hugely bullish on this idea, was not able to join the call today, but can we please see a detailed liquidity model/discussion re how this product does work and scale - and beat it up - before it goes to IIP?

Apologies for not being able to make the call and lacking context on what was discussed there.

Cc @JosephKnecht @overanalyser @mel.eth.

If such content was discussed and shared on the call, I’d love to see it. Ie please share. Thanks so much.

I feel some risk of running at this hard and failing due to liquidity and market structure.

2 Likes

Hi Dev,

Thanks for your bullishness. I share your concern that the lack of liquidity is a major challenge. I discussed some of the workarounds in today’s community call (slides, recording) but they’re not without their own issues.

Does the Coop have any tools or best practices for liquidity analysis? I’m happy to work-up my own but I’d prefer to avoid reinventing the wheel.

Best,
Joseph

2 Likes

No reason. I’m happy with either.

2 Likes

Great deck @JosephKnecht ,

Thank for sharing - and for the fast reply.

Hey, look - if PWG and EWG experts sign off on this - with all their peta brains - I’m cool. There’s a ton of liquidity and execution skills there. And you’ve clearly spent a lot of timing thinking about this and built up huge skills.

I guess a question we might have: we have caps on FLI products now, maybe we’d have one on this too - to manage carefully vs liquidity and market structure/changes to market structure as we grow.

I’m dying for such a product concept.

2 Likes

Maybe OA or another will share some details here… :pray: