Elastic futures (non-technical)

--- title: Elastic futures (non-technical) author: "Prevert" date: "12/02/2022" description: Non technical tags: Elastic --- # Elastic Futures *This is a non-technical presentation of the design. For a more formal approach, see [here](https://hackmd.io/@pre-vert/elastic_futures_technical).* >The common truth about leveraged ETFs is they are just terribly designed financial products that would result in severe loss in the long run. However, perhaps the decentralized version could challenge the design and reinvent the product to be able to withstand the decaying effect. [Coingecko](https://www.coingecko.com/research/publications/part-2-deep-dive-into-decentralized-leveraged-etfs) Leverage is a key building block of financial systems. The huge volume made by CEX and on-chain [perpetual futures](https://insights.deribit.com/market-research/the-quest-for-perp-amms/) is evidence of a strong demand willing to speculate on asset prices. However, while a rich set of leverage solutions are proposed by the markets, they are complex to manage and/or not fit for long-term holding. As most leverage products are based on borrowing, users have to regularly check their funding cost and monitor their margin to avoid liquidation. They need a comprehensive understanding of the risk they take and how to manage it. This complexity led to the creation of leveraged ETFs in 2006 which manage the financial engineering on behalf of their investors. In decentralized finance, the [ETH2x-FLI](https://medium.com/indexcoop/introducing-the-flexible-leverage-index-fli-by-pulse-inc-a369bd422ef) leverages a collateralized debt position by targeting a long 2x exposure to ETH. Despite being a laudable initiative, the design is flawed by the presence of [volatility decay](https://indexcoop.com/blog/fli-volatility-drift) which prevents users from holding the fund over extended time intervals. This note presents an innovative financial arrangement called elastic futures allowing traders to take leveraged long and short positions against each others. Positions have a common forward price as in futures contracts but no preset settlement date. Positions are liquid and can be redeemed against the pool at any time. The absence of duration makes them comparable to perpetual futures with the notable difference that users do not face liquidation nor pay the other side a funding rate. Elastic Futures depart from order book-based (or peer-to-peer) leverage protocols like Dydx and AMM-based (or peer-to-pool) protocols such as GMX, by designing pool-to-pool interactions as in [Mycelium](https://mycelium.xyz/) [perpetual pools](https://mycelium.xyz/blog/leverage-trading) in which the collateral of the long and short positions are deposited in permissionless pools. A major difference is the absence of rebalancing between the pools, a feature which protects the return from [volatility decay](https://www.investopedia.com/articles/investing/121515/why-3x-etfs-are-riskier-you-think.asp). The Table compares the features of futures, Perp, Mycelium's perpetual pools and elastic futures. | | Futures | Perp | Perpetual pools | Elastic futures | | ----------------- |:-------------------------------:|:-------------------------------:|:------------------------------:|:-------------------------------:| | Expiry date | yes | no | no | no | | Funding rate | no | yes | no | no | | Margin management | yes | yes | no | no | | Liquidation risk | yes | yes | no | no | | Volatility decay | yes | yes | yes | no | | Model | Peer-to-peer | Peer-to-peer/pool | Pool-to-pool | Pool-to-pool | Overall, the absence of liquidation, funding rate and volatility decay make elastic futures easy to use and fit for prolonged holding. ## Contract design Traders deposit a collateral asset (typically stablecoins) in a long pool, which offers a leveraged long position, or a short pool, which provides a leveraged short position. Both positions are expressed with respect to a derivative with an accessible price feed denoted $p_t$ (e.g. ETH). A pool is composed of the collateral deposited by users to which is added or subtracted a PNL. The PNL is increasing with the price in the long pool and decreasing with the price in the short pool. Both returns are amplified by a leverage ratio.  Traders who withdraw their funds pay a flat fee reinvested into the smallest pool. Fees paid by the largest pools are amplified by a factor equal to the imbalance ratio once transferred to the smallest pool. The larger the imbalance, the greater the amplification and incentives given to depositors in the smallest pool. ### Example Let us consider the example of a futures contract with a leverage $\times 3$ and a forward price of $100$. The two pools are collateralized with $1$ unit of asset each, so that there is no imbalance. ![](https://i.imgur.com/isb00GR.png =460x285) In the graph, the blue curve is the value of the long pool and the red line the value of the short pool given the price on the horizontal axis. Each pool has $1$ circulating share token so that the lines also indicate the price of the share tokens. At a market price of $110$, a trader may go long and mint long share tokens at price $130$. Days or weeks later, he exits his position at market price $125$ by burning his long share tokens at price $175$. His return rate is $41\%$ given a price appreciation of 13.6%. ![](https://i.imgur.com/nXWcNw1.png =480x300) ### Global settlement If the price moves far away from the forward price, the loss of the losing pool may exceed its collateral. This triggers an onchain settlement process. ![](https://i.imgur.com/kgg1VAc.png =540x260) The trading price interval is $[66,134]$. If the price falls below $66$, the debt of the long pool exceeds its collateral and its value is zero. The value of the short pool cannot exceed $200$ as a result. For the same reason, when the price goes above $134$, the value of the short pool is zero and the value of the long pool reaches a maximum of $200$. A settlement process entails the following smart-contract actions: - the settlement price is frozen - share tokens of the losing pool are worth zero - deposits are disabled in the two pools Withdrawals from the winning pool are still allowed. As holders of share tokens of the winning pool have proportional rights to the collateral in the losing pool, withdrawals have the effect of emptying the remaining collateral in the two pools. The settlement regime ends and the trading regime starts again when the price returns to its trading interval. ## Economic analysis ### Market equilibrium The previous example assumes that the two pools have equal amounts of collateral. In practice, the market is often skewed towards one direction. If traders are bullish, the long pool will be the biggest one. If they are bearish, it will be the short pool. What matters here is pools' respective collateral, not their value, which serve a different function. For example, the pool's value may be low due to unfavorable past price action, yet its collateral can be sizeable if traders expect a strong reversal. In perpetual futures and lending protocols, adjustments of the funding rate guarantee the equilibrium between supply and demand of leverage. In elastic futures, the fee structure and the actual leverage serve a similar function. Withdrawal fees reinvested in the smallest pool incentivize inflows. Moreover, the smaller the relative of size of the pool, the greater the fees compared to the pool's size. In case financial incentives are not effective enough, the leverage of the largest pool will decrease until an equilibrium is eventually reached. Suppose for instance that the market is bullish and that the long pool size is $1.5$ times the size of the short pool. The actual leverage of the long pool is $2$, as shown in the graph (blue line). ![](https://i.imgur.com/E1uV4dn.png =470x320) The more TVL in the long pool, the less long traders benefit from the upside, which was why more TVL was accumulated in the pool in the first place. This cools down the demand for long positions, until an equilibrium is reached. A symmetric mechanism operates when the market is bearish. The graph presents the returns for a long pool size equal to $2/3$ of the short pool size. ![](https://i.imgur.com/0HbTgVs.png =470x320) The leverage of the largest pool declines until part of the demand for collateral is discouraged and remaining traders are still satisfied with the new value. It is as if depositors in the smallest pool made available a fixed aggregate quantity of leverage to be shared between depositors in the largest pool. The equilibrium mechanism is illustrated in the Figure in which the maximum leverage factor is set to 3. ![](https://i.imgur.com/gVw0CGd.png =550x250) The demand for long and short is measured by the amount of collateral deposited in the pools. If the demand for long equals the demand for short, both benefit from the maximum levereage factor equal to 3. If the demand for long leverage is twice the demand for short leverage, long traders benefit from a leverage ratio of 2 instead of 3 and up to 1 if it is three times the demand for short. ### Border effects In borrowing-based leverage products, positions can be liquidated to prevent borrowers from going bankrupt. To avoid costly liquidation, users need to provide additional collateral when markets are volatile. In elastic futures, pools are collateralized for a wide range of price variations. Since the liability of the loosing pool is limited to its collateral, depositors in the winning pool have their profits capped. ![](https://i.imgur.com/S91Llns.png =540x260) For opposite reasons, traders in the loosing pool face limited downside and full upside. They have strong incentives to gamble their losses and bet on a recovery. As a result, traders on the winning side may wish to withdraw their fund before the collateral in the opposite pool is exhausted, whereas traders in the loosing side have no reasons to withdraw and may even deposit additional collateral. A nice property of the model is that progressive withdrawals by the winning side widens the price range over which the futures is traded, which postpones the settlement phase and extends the futures lifetime. Let us see what may happen when the price approaches its upper trading limit. Long traders are incentivized to withdraw their fund, contrary to short traders. Pools' collateral will be increasingly imbalanced in favor of the short side. The animated graph shows what happens when the collateral is more and more concentrated in the short pool, starting from no imbalance. ![](https://i.imgur.com/zVQDJti.gif =470x310) The increasing imbalance between the pools produces beneficial consequences for both sides. First, the short pool becomes more collateralized compared to the long pool, which makes it less likey to go to settlement. In the graph, the settlement price which was $134$ with balanced pools progressively increases to $200$. This gives short traders some breathing room and more chance to recover their losses. They also benefit from a higher pool's value. Second, a higher settlement price means that long traders benefit from an extended upside. The pool's value against which traders can redeem their share tokens doubles in the example (from $200$ to $400$). This reduces their incentives to withdraw early and extends the derivatives' lifetime. A symmetric mechanism takes place when the price is getting closer to the threshold below which the long pool's equity is zero. The graph illustrates this case under the assumption that the size of the long pool is eventually three times larger than the size of the short pool. ![](https://i.imgur.com/X9SSHMq.gif =470x310) An increasing disequilibrium between the two pools extends the price range over which the derivatives can be traded. The settlement price which was $66$ when the pools were balanced, falls to zero for a long pool $3$ times bigger than the short pool. Notice that the settlement price could potentially extends to 0, meaning that a settlement phase never happens on the downside. ### Adding limit orders Although the payoff shapes of long and short traders resemble the shapes of call and put options without a premium, they are different in nature. Traders always have the ability to exit the pool before the price exits the trading interval and the payoff reaches a plateau. Limit sell and buy orders are easily implemented on top of the protocol. In the graphical example, long traders post a sell order at 130 and a buy order at 80. Short traders post a sell order at 70 and a buy order at 120. ![](https://i.imgur.com/VyyDJ6M.png =490x310) Arbitrage bots may compete to take the order whenever they are profitable by either taking the sell order and redeeming the share tokens against the pool or by minting share tokens and taking the buy orders. ### Adding contracts As outlined, the price interval over which users can trade leverage tokens widens when profitable positions are closed. However, this process cannot operate indefinitely. As the pools become more and more imbalanced, the delta of the loosing pool approaches zero, eliminating most incentives to stay. Also, recall that the winning pool earns growing fee income as their relative size shrinks, which guarantees that some collateral will also stay in the winning pool. If the price keeps increasing, the imbalance ratio will stabilize eventually and the contract will enter a settlement phase. Given those considerations, the protocol will gradually open new futures contracts with updated forward prices. As an illustration assume that the price is moving upward and approaching the high settlement threshold. A new contract is then deployed with a higher forward price. As an illustration, the graph shows the value of fours pools, two with a forward price of $100$ and two with a forward price of $180$. ![](https://i.imgur.com/wTBUhT3.png =470x320) This opens up a new set of strategies. Long traders can extend their longs with a higher settlement price (in the example from $134$ to $239$). Or they can take an opposite bet and open a short either in the first contract or the new one.  *See [here](https://hackmd.io/@pre-vert/elastic_futures_technical) for a more formal approach.* <!-- ## Conclusion - The absence of liquidation, funding rate and volatility decay make elastic futures easy to use and adapted to prolonged holding. Contrary to Perp products in which the excess demand for leverage is priced out by the increase of the funding rate, demand for leverage in elastic futures is regulated by the rationing of the side of the market in excess. The equilibrium between the demand for long and short positions is assured by endogenous leverage which plays the role of the funding rate in Perps. | | Futures | Perp | Perpetual pools | Elastic futures | | ----------------- |:-------------------------------:|:-------------------------------:|:------------------------------:|:-------------------------------:| | Expiry date | yes | no | no | no | | Funding rate | no | yes | no | no | | Margin management | yes | yes | no | no | | Liquidation risk | yes | yes | no | no | | Volatility decay | yes | yes | yes | no | | Limit orders | no | yes | no | yes | | Model | Peer-to-peer | Peer-to-peer/pool | Pool-to-pool | Pool-to-pool |