IIP-60: Launching Llama Diversified Index (LDI)

[Llama]

Status: DG1
Author: Llama
Created: 2021-07-06
Requires: IIP-58: Launching Pulse Aggregate Yield (PAY)

Simple Summary

Llama would like to partner with Index Coop (IC) to co-manage the Llama Diversified Index strategy. The Llama Diversified Index (LDI) is a productive and diversified crypto index that optimizes returns relative to risk across various sub-asset classes within the crypto ecosystem.

LDI has already secured approximately $3.5M of external funding. Llama is acutely aware of limited dev resourcing within DeFi and is willing to share the dev workload associated with developing and maintaining LDI. Llama would also like to highlight the 15% allocation to the Pulse Aggregate Yield index (PAY) within LDI.

Llama is truly excited by the prospect of partnering with Index Coop and looks forward to developing a long-term prosperous relationship.

Motivation

This new set has been developed to provide holders a diversified portfolio of productive assets wherever possible to maximise returns. By providing holders with an automated diversified asset allocation solution, LDI enables passive investment in a diversified set that provides access to the main thematic investment sectors on the Ethereum network. The diversified portfolio construction has already proven to have strong product market fit as Gitcoin has expressed interest in becoming an early investor in the product.

In time, LDI will be integrated across DeFi, allowing for extrinsically productivite opportunities to be generated on top of a balanced index, further enhancing the appeal of holding a diversified portfolio. Llama has built an extensive network across DeFi and is able to accelerate LDI integrations across the ecosystem. At the time of writing, Llama has approximately 25 contributors that are also members of other DeFi communities.

LDI has several key advantages over traditional diversified portfolios:

• Utilization of productive assets where possible.
• A more sophisticated asset allocation methodology that optimizes for Sharpe ratio while integrating appropriate constraints to ensure diversification among established crypto sub-asset classes.
• Low vol USD yield aggregation (PAY)
• Access to a liquid secondary DEX market
• LDI productive use cases

Llama has already received expressions of interest from various parties regarding this particular product and intends to create complementary products as part of our overall treasury management strategy solutions.

Size of Opportunity

DeFi treasuries have amassed a total AUM of over $6B. As DAOs seek to diversify their treasuries, passive and diversified portfolio solutions will enable simplified management. The total addressable market for a diversified Ethereum-focused crypto portfolio is tremendous. Capturing a small portion of this overall market means generating significant AUM for Index Coop. A diversified crypto index product currently does not exist within DeFi and therefore represents a unique opportunity for Index Coop to partner with a Treasury Management specialist in Llama.

Llama’s network includes university endowment funds which Llama believes will find LDI attractive. Llama has already discussed the LDI with a number of DAOs and found strong support for the product which can be used as an allocation within a broader treasury management strategy.

Product Differentiation

There are currently no diversified portfolio offerings available as crypto-native structured products that optimize for returns relative to standard deviations among various sub-asset classes. Indices like BED offer simplified solutions for the risk assets within a portfolio, but they do not allocate to stablecoins (an essential component of a diversified crypto strategy). In addition, the methodology of BED equal-weights the sub-asset classes (e.g. Bitcoin, ETH, and DeFi) rather than optimizes risk/reward between them. While there is definitely a market for equal-weighted indices like this, the LDI intends to offer a more sophisticated portfolio management solution.

The introduction of a stablecoin allocation will lead to automatic rebalancing toward risk-on assets during periods of market stress and toward risk-off assets during periods of market euphoria. This diversified approach ensures contrarianism and is a key component of a portfolio management mindset.

Finally, a bias toward productive assets addresses a key treasury diversification goal. For example, Nexus Mutual recently invested a large amount into stETH and Fei Protocol recently approved purchasing stETH. Since this product attempts to create a diversified allocation while also ensuring productivity where possible, there is no need for a DAO treasury seeking a simple, passive solution to do more with their investment assets outside of holding LDI.

Liquidity Analysis

The underlying assets within LDI are all readily available within DeFi as ERC20 tokens. Most tokens can be purchased via Sushiswap, Uniswap, Curve by depositing capital into the respective protocol. For tokens traded via DEXs, there is ample liquidity to facilitate rebalancing when compared to some of the more illiquid components in DPI.

The only token with underlying liquidity concerns is the productive iteration of AAVE. stkAAVE has a lock up period and xToken products like xAAVEa & xAAVEb also rely on AAVE being unstaked periodically to provide unwrapping liquidity. Llama expects at the time of launch xToken will have launched AAVE-xAAVEa & AAVE-xAAVEb pools on Uniswap V3 which are expected to provide sufficient rebalancing liquidity. These contracts are currently being audited.

LDI intends to utilise a combination of Exchange Issuance and on market liquidity to facilitate rebalancing the PAY allocation within LDI.

Methodology

LDI is intended to provide a diversified portfolio of crypto assets across a variety of sub-asset classes. The Index strives to provide an effective portfolio solution for DAO treasuries and individuals alike.

The allocation will initially be divided among 6 sub-asset classes: Stablecoins, Bitcoin, ETH, DeFi, Metaverse, and Web3 Infrastructure. These sub-asset classes were determined by analyzing the current crypto market to identify key themes and differentiated asset types. Sub-asset classes could change over time as new themes and asset types emerge. The reasoning for the current sub-asset class elections are as follows:

• Stablecoins (Target: 15%) - Stablecoins are largely uncorrelated to the broader crypto ecosystem. The stability inherent in stablecoins make them essentially a cash equivalent while offering returns comparable to traditional equities when deployed productively. Having a sizable stablecoin allocation allows for a contrarian investment approach during the rebalancing process: risky assets are trimmed during periods of strong performance and are added to during drawdowns. This rebalancing approach has been shown to outperform the classic buy-and-hold strategy over multiple market cycles.
• Bitcoin (Target: 15%) - Bitcoin has the lowest volatility of the other risky asset types included in the index, serving to bring down total portfolio volatility without sacrificing too much upside. Its fixed supply and position as the largest high-quality crypto asset makes it a kind of safe-haven asset during periods of market instability while displaying characteristics reminiscent of traditional real assets like gold.
• Ether (Target: 35%) - Ether is the life-blood of the Ethereum economy and the main asset used to conduct operations by DAOs outside of their native governance tokens. Since ETH is essentially the money of the ecosystem, it is a primary asset for DAO and individual portfolios. Investing in ETH provides exposure to a variety of crypto themes while also providing a remarkable risk/reward profile versus other crypto assets including Bitcoin.
• DeFi (Target: 20%) - DeFi has emerged as one of the largest opportunities within crypto, and is arguably one of the first killer applications of blockchain technology. There is strong demand for DeFi assets as evidenced by new centralized offerings (e.g., the Bitwise DeFi Index), the success of DPI, and the inclusion of DeFi in the BED meme. In addition, many DAO treasuries have shown interest in acquiring the assets of other DeFi protocols.
• Metaverse (Target: 10%) - The Metaverse is quickly emerging as one of the key themes of the 21st century as we spend an increasingly larger amount of time in digital spaces. The theme has been highlighted in traditional finance, specifically around the Roblox IPO, but NFTs, blockchain gaming, and crypto in general provides a better platform for some of the core tenets of the theme. While the opportunity is still in the early innings, the success in capturing the imaginations of investors thus far warrants an investment in the space.
• Web3 Infrastructure (Target: 5%) - Web3 Infrastructure, including oracles and sub-graphs, have enabled DeFi to flourish. As a sub-asset class, it is expected that middleware remains less correlated to other sub-asset classes like DeFi and the Metaverse as it is more directly tied to the growth of on-chain data.

Llama will periodically review the strategic weightings above to determine if they are still suitable based on market dynamics. The allocation size of each sub-asset class will be determined based on qualitative and quantitative characteristics, including but not limited to potential size of the opportunity, correlation to existing asset classes, investor appetite, long-term viability, and suitability within a treasury portfolio. Llama will maintain a committee within LlamaDAO focused on analyzing current and future sub-asset classes.

The allocation to the underlying constituents of each sub-asset class will follow a risk-optimized market cap weighting methodology outlined below:

• Llama will use a combination of quantitative and qualitative screeners to identify core assets of each sub-asset class to be included. The selection process will seek to identify best-in-class assets that are representative of the sector.
• A fully-diluted market-cap weighted methodology will be used on the screened assets to determine an initial weight. A ‘10.0% of sub-asset class’ band will then be applied around the initial investment allocation to determine the possible tactical allocation range. The maximum allocation for each asset will be limited to 5.0% of the total portfolio to ensure diversification. The minimum allocation will be set at 1.0% of the total portfolio to ensure inclusion justifies the frictional costs associated with rebalancing.
• The final allocation is determined by optimizing for maximum Sharpe ratio (return relative to unit of risk). Better risk/reward assets will be weighted more heavily, while less capital efficient assets will be underweighted relative to the initial market cap weighting within the defined band.
• No leverage or shorting may be used.

Within each sub-asset class, there will be a bias toward productive iterations of each asset.

Sample Composition

A sample portfolio composition is shown below. This is subject to change and shall be confirmed prior to launch.

Note:

1. TokenSet adapters are shown to indicate relative ease of creating the LDI set
2. Issue contracts will require users to deposit wrapped tokens in the correct portions
3. Exchange issue function is required
4. Exchange issuance shall automate multiple purchases of underlying tokens, minimising slippage and accommodating suitable large trades that would otherwise cause significant slippage in the underlying tokens
5. Redemption will release the underlying tokens in the differing proportions
6. Productive iterations of underlying assets can vary with time based to optimise returns

Maintenance

Llama will periodically provide Index Coop with an updated LDI composition. Llama is also willing to discuss providing dev resources to assist with the rebalancing process. This can also be extended to include developing any additional adapters that may be required. The prospect of partnering and potentially sharing developer resources symbolises Llama’s commitment to building an alliance within DeFi. LDI = Let’s Do It!

Initial rebalancing frequency is expected to be quarterly. Llama seeks the ability for DAO treasuries to deposit large sums of capital into a contract that acquires the underlying assets gradually over time to minimise slippage. Llama would also like the same functionality built into the redeem function to allow large positions to be unwound over time.

Fee Structure

For Llama and Index Coop to discuss.

Author Background

Llama is building primitives for on-chain treasury management. Llama is working with Aave, Gitcoin, Uniswap, Radicle, ARCx, FWB, and has other top tier DeFi integrations in the pipeline. Llama aims to set the gold standard for DAO treasury management and create category defining templates that apply across DAOs.

Copyright

Copyright and related rights waived via CC0 4.

[overanalyser]

Welcome @Llama

I must say that this is a most interesting proposal, It is certainly an target market of interest and the portfolio of underlying tokens and weights looks interesting.

Thank you.

[BigSky7]

Wow - great work llama team. Echo @overanalyser that this is one of the most interesting methodologies we’ve seen so far.

I can’t wait to dive into this proposal this weekend.

Let’s f*** Do It

[Kweiss]

Kyle from Gitcoin here - I just want to chime in and mention that the Llamas team is always great to work with and we would be excited to move some of our treasury to an index like this (organization treasury… DAO treasury is up to the community).

Giving communities some exposure to more volatile assets, while still offering a low and mid volatility mix in the index is highly attractive. I suspect many other DAOs and larger organizations would use an index like this as a “prudent” treasury management option.

I appreciate the Llamas crew for suggesting, and for this community for considering

[qlx]

Good work everyone from Llama.

I think this is an interesting proposal. The focus on productive assets across a variety of classes looks potentially fruitful and distinct within the current Index Coop portfolio.

Did you ever consider including MVI and the forthcoming DATA index for the metaverse and data categories? I am wondering what the merits of the proposed tokens in these categories are over those aforementioned indices. Perhaps there is the intention to put the tokens from those two categories to work as well? I can appreciate that using those two indices would detract from the productivity intent for LDI. Furthermore, how confident can we be of a sustained period of productivity for metaverse tokens? Perhaps this is dealt with through rebalancing. Would there be a consistent productivity function across this category? For example, would this be staking only or would interactive productivity be considered (i.e. using tokens in games)?

Good proposal, looking forward to watching it develop.

[Thomas_Hepner]

@Llama is building a strong reputation in the DAO ecosystem as a Treasury Management service provider so I am very excited to see this proposal put forward. In my view, there is great potential for a long-term win-win relationship between Llama and Index Cooperative.

I am incredibly encouraged by the overlap between the two communities. There is a lot of cross-DAO engagement from contributors like @HelloShreyas, @Matthew_Graham, @verto0912 , @Kiba , @DOC , and others that I am assuredly missing.

I am sure there will be many iterations to this proposal, but I am strongly in favor of Llama becoming a Methodologist with the Index Cooperative.

Here are a few questions and comments on the details of this proposal:

1. Could you elaborate on what this means given that LDI does not exist yet?

@Llama

LDI has already secured approximately $3.5M of external funding.

2. How are the targets for each asset class within LDI being determined? It is very strange to me that Ether has 2x the target allocation as Bitcoin and Metaverse 2x the allocation of Web3 Infrastructure given that Bitcoin has ~2.5x Ether’s market capitalization and Web3 Instrstructure has ~5x the market cap of the Metaverse.

3. Has Llama considered changing the name to reflect that this product is intended to be a Treasury Management solution? For instance, is this product mostly intended for DAO treasuries and institutions or is it also intended to be a product for individual investors?

4. @Matthew_Graham @prairiefi @ElliottWatts Is this a product Index Cooperative should consider for our Community Treasury? If so, in what size?

5. I found it strange that PAY is the included in the proposal, but not DPI, MVI, or DATA. Why leverage PAY and not the others? Could Llama provide additional context on this consideration?

6. Comment on Intrinsic Productivity and “dogfooding” Index Coop products: DPI, MVI, and DATA would all greatly benefit from Intrinsic Productivity (IP) and this has been discussed at length at Index Cooperative. @Kiba suggested xSUSHI and yvYFI be used to Activate DPI Intrinsic Productivity, but those discussions were put on hold after disagreements in the Index Coop community. It seems like there is an opportunity to build out IP for those products first, and then incorporate those products in LDI. Thoughts?

7. Timeline: When is Llama hoping to launch this product by? Is Llama hoping or expecting to receive INDEX rewards from the Methodologist Program? I want to understand Llama’s expectations of the Index Cooperative relationship.

[prairiefi]

@Thomas_Hepner

Is this a product Index Cooperative should consider for our Community Treasury? If so, in what size?

++ @AcceleratedCapital We intend to have an IIP within a few days that outlines our initial recommendation on productive stablecoins within the treasury and how the Operations Account and Investment Account would have different constraints. The Operations Account would focus on minimizing layers of smart contract risk, and the Investment Account would be more enterprising / have more latitude in search of total return, understanding it has a longer horizon. LDI could potentially be an investment account option, but not operations account with the way we’ve drafted the proposal. More to come, and we welcome the discussion.

I interpret a product like LDI as an off-the-shelf investment account, but may not be aligned to our exact goals even if it is a good solution for others. This is a discussion we’ve also had with SYI/PAY, and we can be supportive of the product even if it isn’t our predominant treasury diversification method.

[Llama]

Hi @Thomas_Hepner

Thank you for the feedback. We will try to answer all of these questions as best we can.

Responses:

1. Llama is already in discussions with various entities about LDI. Although these discussions are at early stages, Llama has identified investors willing to invest into LDI.
   Llama is excited and looking forward to working with Index Coop’s BD team to help grow LDI as we believe the initial capital is a good indication of product market fit.

2. LDI is not a market cap weighted product. LDI contains tactical allocation and risk optimised return driven position sizing.

3. The name of the product will include Llama for branding and marketing purposes. The Llama brand will be linked to the success of this product much the same way as DeFi Pulse is linked to the success of DPI. LDI is a Diversified product and is appealing to all types of investors. Llama believes partnering with Index Coop will further strength the branding around LDI and also broaden the reach of the product. There is no desire to change the name.

4. NA

5. PAY is a more complicated product relative to DPI, MVI or DATA and harder to replicate. PAY is an ERC20 token and all of the compounding benefits are built into the price - these are ideal qualities for inclusion in LDI. However, Llama can reassign the cash allocation within LDI to a different asset.
   Llama did look at including DPI, MVI and DATA. Our analysis indicated the risk optimised return analysis methodology leads to higher returns over time.

6. The LDI design utilises productive iterations wherever possible. Similar to the overall tactical allocation, this is by design and intended to be a feature of the product.

7. Llama would like to have LDI trading within 6-8 weeks. Llama is looking forward to entering commercial discussions after DG1.

Llama would like to develop an open, transparent and mutually beneficial relationship with Index Coop community. Llama has strong working relationships across DeFi which will help support distribution efforts and integrating LDI into other protocols. Llama is also able to assist with rebalancing, integrations with Aave and Compound etc… which require developer lift. In a broader context, Llama is able to share the developer workload needed to support the product.

[Thomas_Hepner]

@Llama Thank you for the thoughtful response to my questions.

A few quick comments / questions:

@Llama

The Llama brand will be linked to the success of this product much the same way as DeFi Pulse is linked to the success of DPI. LDI is a Diversified product and is appealing to all types of investors.

Definitely agree that “Llama” should be included on the name. I was pondering if having the word “Treasury” in the product name was something Llama was considering for product positioning.

@Llama

Llama did look at including DPI, MVI and DATA. Our analysis indicated the risk optimised return analysis methodology leads to higher returns over time.

Did the return analysis consider if these products as is or their intrinsically productive versions?

@Llama

Llama would like to have LDI trading within 6-8 weeks.

Llama is hoping for a mid-September or early October launch? I would love to have LDI launched that quickly, but think this is extremely aggressive timing given that it is (1) dependent on launch of PAY, (2) engineering resources are constrained, (3) the community has not come to consensus on how to safely support yield generating products, and (4) LDI needs to be prioritized vs other products in the pipeline.

@overanalyser @puniaviision @dylan Any thoughts on feasibility of this timeline?

[Monportefeuille]

Interesting proposal indeed team @Llama, thanks for putting it all together, well aware and respectful of the amount of work and focus this represents.

I will echo @qlx and @Thomas_Hepner in saying I also couldn’t help thinking how this methodology could integrate with existing and potentially forthcoming Index Coop products : MVI, DATA or even iRobot that we also have just submitted as an IIP.

For reference, iRobot is the formalization of a proposal we originally put up for discussion under the name PMI. Its goal is to help individuals, notably those with little experience / time or appetite for portfolio management, to automatically select and rebalance the components of a sector-agnostic token basket based on their risk-adjusted returns.

While the approach and the main target customer of LDI might be different, instead of opposing the 2 methodologies / products it seems worth reflecting how they could mutually benefit from each other - and more generally how this could be the case for Index Coop’s broader palette.

For example, the Robot Index uses the annualized Sortino ratio as key performance indicator, mainly for 2 reasons :

• The Sharpe ratio does not differentiate harmful volatility from overall volatility - which we found out makes it a less accurate performance differentiator between crypto-assets.

• With price histories of different lengths, to be more accurate it also makes sense to normalize the period of time over which the ratio is calculated.

Furthermore a quarterly rebalancing cadence seems to be quite long to capture the changes in market dynamics driving the evolution of these indicators.

Finally, my understanding (but please correct any mistake) is that the initial market cap weighting combined with the maximum 5% of total portfolio allocation per asset limit the effect of the final Sharpe ratio optimization.

Again, I’m not taking a deliberately critical stance here - which would probably be pointless - but rather trying to dig out any constructive and mutually beneficial synergies for the Coop.

For example, a backtest of the proposed sample composition against an option integrating productive assets whenever possible, MVI, DATA and iRobot would be quite interesting - happy to provide performance data from the last 3 months, or anything you may require for iRobot !

[puniaviision]

@Thomas_Hepner

@overanalyser @puniaviision @dylan Any thoughts on feasibility of this timeline?

I agree with your assessment that this is likely an aggressive timeline given the dependency on PAY, which is a complex product to launch that will require a lot of engineering work.

[mason_c]

I was lucky to give feedback on this product as it was being developed.

I think it could be a useful product for many treasuries, the BadgerDAO treasury certainly included. Per naming suggestions above maybe it gets called the Llama Indexed Treasury, or another play on words but I do think the clarity of the word treasury in the name is a good point.

The allocations to productive assets are in my mind quite important as mentioned above for cash flow reasons. I can even imagine a way in which the Index could be leveraged and assets within the pool rented out for utility purposes, a distant but relative example would be related to metaverse assets.

The questions in my mind are related to others above, how can we leverage and dogwood existing index products to create deeper liquidity and use cases?

It stands to reason that the model could perhaps shift allocations between the already existing index products and direct productive allocations. This may provide an interesting way to shift risk a bit more dynamically.

I look forward to reading further discussion on this!

[kermankohli]

Super interesting proposal and seems complimentary to work done with existing DAOs at Llama such as ARCx. Happy to provide any feedback as an external stakeholder that may be interested in this product.

[Llama]

Hi All,

Thank you for all the feedback on the proposal.

We would like to share some further feedback received on the proposal and request for an IIP number to be assigned along with creating a discord chat for LDI. Llama is excited to be working with Index Coop and looking forward to DG1. (@anon10525910 and/or @Pepperoni_Joe)

As alluded to in the post above, there is capital awaiting the creation of LDI and we are keen to move quickly with the proposal. The last tweet on the thread linked below is particularly exciting, Let’s Do It.

DefiDollar

[verto0912]

@Llama

As alluded to in the post above, there is capital awaiting the creation of LDI and we are keen to move quickly with the proposal. The last tweet on the thread linked below is particularly exciting, Let’s Do It .

https://twitter.com/defidollar/status/1417403517111730177?s=20

I think it’s premature to push this to DG1 given that PAY hasn’t gone through DG1 yet and this product is going to be behind PAY in the pipeline in terms of execution. With DATA, PAY and inverse FLI, I see our capacity to process more products right now as limited. See engineering and product constraints.

Further, we used to have a requirement/tradition for a community call before the DG1 vote. We have strayed from that recently and I would like to see it re-introduced.

[Thomas_Hepner]

@verto0912

we used to have a requirement/tradition for a community call before the DG1 vote. We have strayed from that recently and I would like to see it re-introduced.

@verto0912 Would you like to see this for DATA ahead of DG2 given we did not do for DG1? Do you think the community would value that?

[anon10525910]

@Thomas_Hepner

Do you think the community would value that?

Personally, I have found these calls very useful/valuable. Important questions get asked & answered.

[anon10525910]

as @Pepperoni_Joe is offline for a bit, I reviewed against IIP-2 Index Addition Proposal Structure and have added the IIP #: 60.

Creating a discord channel next

[EDIT]: I have also added the-below to the header, @Llama please correct if I am mistaken

Requires: IIP-58: Launching Pulse Aggregate Yield (PAY)

[Llama]

Hi @anon10525910,

Can we please set up a community call for the 21st July at 9am PST time between the Llama community and the Index Coop Community.

If questions can be shared prior, that would be greatly appreciated and enable us to provide the most informative answers.

We would then like to have the DG1 vote scheduled to start later that same day on the 21st July.

Ideally we would like to include PAY but it can substituted out if required to accelerate the path to market.

Looking forward to working with the Index Coop.

[anon10525910]

@Llama

21st July

Happy to get a community call on the books! However, that date (July 21) is in the past.

Could you choose a date in the future?

Ideally, there is enough lead time to collect questions and prep. For example, BED Community Call - Announcement & Request for Questions