Launching the next wave of FLIs

[Jo_K]

IIP: XX , XX+1, XX+2, XX+3
Title: Launching new FLIs
Status: draft
Author: Pulse Inc
Created:

Summary

In view of the community’s excitement for new leveraged products and since the Set Protocol team has also finally merged the Aave leverage module last week, Pulse.inc would like to propose that the Index Coop manages a new set using the FLI strategy proposed in IIP-13.

We previously have taken the approach of proposing FLIs one at a time to gauge the market need as well as testing the methodology and its implementation. We chose to post one overall post and will be taking each of these products to DG1 independently.

After we feel that the Inverse FLI has garnered enough confidence through its implementation, we will be proposing similar iFLIs for all the below tokens as well.

We now propose the launch of four new 2x long FLIs for popular high volume tokens in the space:

UNI2x-FLI - One of the highest volume traded leveraged ETFs on centralized exchanges

LINK2x-FLI - Giving the marines a way to trustfully leverage their positions will be an interesting product to say the least

DPI2x-FLI - DPI is moving fast towards being integrated on Aave, and as soon as this is done, we would love to have a FLI set for this

MATIC2x-FLI - Thanks to concentration of liquidity on Aave Polygon market, this set will be the first Polygon based FLI as soon as the infrastructure is ready to support leveraged products on the Polygon network

Abstract

These new sets, if launched, would be based on V0.1 of the FLI (same as both ETH2x-FLI and BTC2x-FLI)

Initial parameters for these new FLIs:

Uniswap Flexible Leverage Index:

• Underlying Asset: UNI
• Target Leverage Ratio: 2
• DeFi Lending Protocol: Compound
• Maximum Leverage Ratio: 2.3
• Minimum Leverage Ratio: 1.7
• Recentering Speed: 5%

Link Flexible Leverage Index:

• Underlying Asset: LINK
• Target Leverage Ratio: 2
• DeFi Lending Protocol: Aave v2
• Maximum Leverage Ratio: 2.3
• Minimum Leverage Ratio: 1.7
• Recentering Speed: 5%

Polygon Flexible Leverage Index:

• Underlying Asset: MATIC
• Target Leverage Ratio: 2
• DeFi Lending Protocol: Aave Polygon
• Maximum Leverage Ratio: 2.3
• Minimum Leverage Ratio: 1.7
• Recentering Speed: 5%

DeFi Pulse Flexible Leverage Index:

• Underlying Asset: DPI
• Target Leverage Ratio: 2
• DeFi Lending Protocol: Aave v2
• Maximum Leverage Ratio: 2.3
• Minimum Leverage Ratio: 1.7
• Recentering Speed: 5%

N.B.: these parameters may be changed/altered before these products reach DG2

We are proposing that these FLI products utilize Circle’s USDCoin (USDC) based on the liquidity & utilization metrics on both Compound & Aave. This may be changed in the future if other stablecoins present sufficient liquidity and utilization on supported lending protocols.

These new FLIs work the same way as ETH2x-FLI & BTC2x-FLI do in the sense that user’s deposited assets are locked in a money market and used as collateral to borrow more of the underlying asset.

Motivation

Flexible Leverage Index makes leverage effortless.

The User would not have to worry about:

• Monitoring his leveraged loan 24/7, having to always be ready to act.
• High fees, transactions not being included fast enough or the relative UIs being unresponsive during times of high volatility.
• Paying for overpriced stablecoins to deleverage on time or panic trading to save his positions.

FLI has several key advantages over Legacy Leveraged Tokens:

• Zero slippage via composable entry and exit.
• The unique index algorithm reduces rebalancing needs by an order of magnitude.
• Emergency deleveraging possible during Black Swan events for additional fund safety.

Size of opportunity

Lending protocols are currently leading the Decentralized Finance space, having multiple Billions in Total Value Locked. There is certainly no shortage in interest for collateralized debt positions, one can simply visit DeFi Pulse and look at lending protocols rank and TVL to get a sense of the magnitude of opportunity this new kind of index can potentially present.

Differentiation

FLI being very dissimilar to any existing or proposed Index opens up a new category of leverage-based Indexes.

We believe that the proposed sets make natural additions to the existing FLI roster and to the IndexCoop asset line-up.

On-chain liquidity analysis

UNI:
Compound: Available liquidity of 403M$ from 2571 suppliers, 25M$ borrowed by 147 addresses
Aave V2: Market size of 122.5M$ with 1.38M$ borrowed

LINK:
Compound: Available 118.9M 575 from suppliers on, 18.4M borrowed by 145 addresses
Aave V2: 563.2M in Market size and total borrowed volume is $ 6.71M.

Matic:
Aave’s Polygon market: 249.2M in Market size and total borrowed volume is 72.2M. Note that this liquidity is only available on the Polygon market, and the MATIC2x-FLI will be the first Polygon-based FLI as soon as Set is ready to launch its leveraged products infrastructure on Polygon.

DPI:
The DPI is moving fast towards being integrated on Aave. Based on the assets current market cap north of 150M, we expect it to have good liquidity on the platform. As soon as the listing happens & liquidity is bootstrapped, we will be analysing liquidity requirements and publishing those for consideration.

Methodology

Objective

Flexible Leverage Index enables market participants to take on leverage while minimizing the transaction costs and risks associated with maintaining collateralized debt.

Definitions

• Borrow Rate — the cost to borrow the asset at the DeFi Lending Protocol over the most recent epoch.
• Epoch Length — the time between rebalances.
• Target Leverage Ratio (TLR) — the long term target for the value of the assets held by the index divided by the value of the debt held by the index.
• Current Leverage Ratio (CLR) — the value of the asset currently held by the index divided by the current value of the debt held by the index.
• Maximum Leverage Ratio (MAXLR) — the highest leverage ratio the index will ever have after a rebalance.
• Minimum Leverage Ratio (MINLR) — the lowest leverage ratio the index will ever have after a rebalance.
• Re-centering Speed (RS) — the rate at which the Current Leverage Ratio is adjusted each period to return to the Target Leverage Ratio, when the index is not being adjusted back to the Maximum Leverage Ratio or the Minimum Leverage Ratio.

Index Price:

FLIt = FLIt-1 * (1 + ((Pricet/Pricet-1–1) * CLRt-1 — (BorrowRatet * (CLRt-1 -1)/CLRt-1)))

Calculation of the new Current Lever Ratio for the period:

CLRt+1 = max(MINLR, min(MAXLR, TLR * (1 — RS) + CLRt * RS))

Fee split

Flexible Leverage Index will have a streaming fee of 1.95% (195 basis points) and a 0.1% minting /redeeming fee. The revenue generated from the streaming fee will be split 40% to DeFi Pulse and 60% to Index Coop.

Author background and commitment

DeFi Pulse and the Pulse Inc brand are committed to maintaining and creating indices. As well as driving the continued growth of the Index Coop.

DeFi Pulse is the leading website for the latest analytics and rankings of DeFi protocols. DeFi Pulse’s rankings track the total value locked into the smart contracts of popular DeFi applications and protocols. Providing key insights and educational content to help more newcomers go from zero to DeFi.

[Matthew_Graham]

Hi @Jo_K

Awesome post When DG1 …

Would it be possible to add the following products:

1. xSUSHI / SUSHI FLI
   Aave V2: Market size of $196.3M with $0.8M borrowed. The native APR would be a nice addition. Interest bearing leverage trading.

2. AAVE FLI
   Aave V2: Market size of $694M with no ability to borrow.

3. Long ETH/BTC FLI
   The flippening trade.

Curiously, are there any ratio products in the pipeline that we can look forward to?

With the increasing FLI product range offering, does DeFi Pulse have like a FLI portfolio manager that can be a point of contact who can actively contribute to the FLI pod ? I think this will really help integrate both teams and enable direct/transparent product specific communication.

[Jo_K]

Those will definitely be proposed after this batch!
we want to start focusing on FLI pairs as well (long/short) once iETH-FLI has had time to mature, but a rough roadmap like the one proposed above is what we hope to be able to launch in the near future with the coop.

These would be amazing products to launch so we are constantly revising and back-testing, trying different permutations to find a way to propose more successful products to market (including ratio FLIs!)

[Jo_K]

about the portfolio manager, the Pulse.inc team is starting to grow and we might have a new addition sometime soon, we’ll definitely be in touch as soon as that happens

[Matthew_Graham]

@Jo_K

we are constantly revising and back-testing, trying different

Hi @Jo_K,

This to me reads like a fantastic opportunity for both Index Coop and DeFi Pulse to work more closely together. There is perhaps some overlap in what each team is working on and we can definitely combine ideas and collaborate together to maximise the FLI series together.

Perhaps someone from DeFi Pulse can join the FLI Pod and we build on past success by further integrating the two teams.

[ABertoG]

Super excited to see this collaboration

It’s great to see Index leverage Aave’s diversity of assets. Would also be great to consider BTC/ETH LP 2xFLI

Can’t wait to try the flippening strategy

[afromac]

Hey Jo. Great to see we are all in agreement that the best path forward is rapidly launching more leveraged products.

Before we commit to this next phase of leveraged products, there are a number of outstanding issues regarding the FLI suite that need to be addressed - both in terms of performance and licensing fee.

Performance Issues

In terms of performance, we have some open questions on how the FLI products behave. The image below shows the relative performance that ETH-FLI has compared to a simple 2x strategy where the product rebalances to a set leverage ratio every 24 hours.

The data that we have points towards the fact that the flexible leverage methodology underperforms a simple leveraged strategy in all markets. As is clear from the data above, holding a simple leveraged position, such as that used by FTX ETHBULL would be both more effective in a bull and in a bear market.

In fact it appears that a FLI methodology will consistently underperform a simple leveraged strategy over time.

This noted underperformance is reflected in the feedback we have gotten from numerous customers over the last few months. If you were to spend a few minutes scrolling through the #fli-discussion channel, you would see this is the prevailing sentiment.

Two takeaways:

1. Our customers are sophisticated investors, and want to understand the product and monitor its performance - this includes the flexible leverage ratio.
2. It has become widely known that the flexible leverage ratio, as it operates today, is ineffective and many of our customers are aware of this.

This leaves me with a number of questions also:

1. How much research and testing was actually done to decide on a flexible leverage ratio?
2. Can you share some of these results with us?
3. We are still operating on V.0 of the FLI methodology. When can we expect V.1 and what will be the improvements?
4. The FLI products have found strong market fit and been profitable for Defi Pulse and the Index Coop. Would it make more sense to use that as an opportunity to refine the methodology before launching more products that will create the same issue for our customers?

With regards to this performance issue, it seems that a more practical approach would be for the Index Coop to begin launching our own line of leveraged products. This way we can provide that product option to our customers while DeFi Pulse continues to refine the FLI methodology. When V.1 is ready, we can implement it on the existing FLI products and decide whether to begin rolling out the FLI suite based on the (assumed) evidence of it’s improved performance.

Licensing Fee Issues

An outstanding issue that has yet to be addressed is the high gas costs associated with managing these products, and the asymmetric risk that creates for Index Coop relative to DeFi Pulse. Quickly looking at Uniswap, it is clear that all of the suggested assets for inclusion as FLI tokens, have significantly less trade volume than BTC.

Asset	Trade Volume
ETH	$1.02B
WBTC	$58.7M
UNI	$14.5M
LINK	$13.6M
MATIC	$8.9M
DPI	$689.9K

This is concerning because BTC2x-FLI has struggled as an asset, and has placed a significant burden on the coop in managing related gas costs.

The BTC-FLI product is still operating at a loss.

We know from historical data that during periods of high volatility, the gas costs of rebalancing the product can be upwards of 1000% of daily revenue. If markets were to suddenly collapse and remain depressed for an extended period, it could take upwards of three years for a FLI product to become profitable again.

My worry stems from the fact that all of the suggested products could plausibly trade at significantly lower volume than BTC2x-FLI but would require similar gas expenses to rebalance during neutral/positive market conditions. There is no guarantee they would be profitable for the Coop to operate.

In the event of a major downturn, IC could be in a position where it has to manage multiple products that are underwater - while still paying a 40% fee to DeFi Pulse before expenses are accounted for. This is not a hypothetical situation. This is exactly the situation we have been in with BTC2x-FLI since launch.

Summary:

Index Coop is now positioned to rapidly deploy a range of leveraged products that will give our customers a broad ecosystem of assets to trade. The opportunity is great, but so are the risks. In order to advance this plan, we need to address the outstanding issues relating to licensing fees that we have been discussing for over a week. These are valid concerns, and we would appreciate a constructive response from DFP.

Similarly, it has emerged that there are deficiencies in the actual performance of the methodology. We appreciate that the products were brought to market quickly and currently exist in a V.0 state; however it would be irresponsible to launch a third, fourth, fifth, and sixth FLI product knowing what we know now about the flexible leverage underperformance. Before we can proceed in good faith, it’s important for the community to know what is being done to address these issues and how DFP intends to improve the methodology. In the meantime, it makes more sense for the Index Coop to launch these products ourselves as we can do so more profitably, and deliver more for our customers.

[scott_lew_is]

hi afromac,

your backtest is incorrect. if you could post your source data i would be happy to show you where.

thanks,
scott.

[afromac]

Hi Scott.

No problem. I will share with you on discord.

[0x_Dev]

What was the outcome of this conversation @afromac ?

[afromac]

Hi Dev

Thanks for your question on this. It has reminded me to follow up on something that I should really have addressed earlier.

In short, @scott_lew_is is correct and my backtest for the simulated ETH2x product is inaccurate:

A couple of important factors that the backtest is missing are:

• Cost of borrowing USDC
• Streaming fee
• Price impact on rebalancing trades

Because of these missing factors, the benchmark that I measured ETH2x-FLI against is unfair. As my model is highly idealized/simplified, and the ETH2x-FLI performance is taken from real data, comparing the two is misleading.

Members of the product team are working on a more accurate model, and that will be presented when it is ready.

To explain my own actions, I acted hastily during a tense period and did not do enough due diligence on the information I was providing. At the time, I believed that the backtest I was producing was simple, but accurate enough for my purposes. I realize now that this is not the case. I apologize for misleading the community, and also to DFP for mischaracterizing the product.

[kman2324]

I don’t know if it’d be possible but it’d nice to see a percentage breakdown of where the ETH2X-FLI fails to match an idealized 2xETH return. For example, what percentage of the “shortage” is due to borrowing fees? What percentage is due to streaming fees? Etc, etc. I know it will vary with time but it might be useful for new potential customers to be told why the ETH2X-FLI doesn’t just match a hypothetical 2xETH expected return.

[0x_Dev]

Hi @afromac,

Thank you for your reply - and that’s honourable and open to respond that way.

I’ll look forward to seeing the new model and its outputs. It would also be great to compare this modeled FLI performance against the fixed leverage ratio product’s potential performance too. In my mind I’m building a comparison matrix of these two products, scored from a client’s perspective.