Elastic Decentralized Loans powered by Chainlink
Introducing vUSD and vETH
Note: This article represents a snapshot of a work in progress for the upcoming rebase decentralized lending platform. All the descriptions in this article are subject to change at anytime. We invite the community to refine these ideas with us to make a better platform.
Digital assets pegged to a stable asset with elastic supplies are a fundamentally new paradigm in the world of decentralized finance. These assets work by pegging a token to a price feed with pre-defined “rebasing” rules. If these assets were securities and not digital commodities, rebasing events, when the supply of the asset has changed, would be equivalent to automatic stock splits or reverse stock splits.
In the absence of external market forces, speculation drives the market mechanics of elastic assets with traders designing strategies based primarily on the market capitalization of the underlying asset. No inherent incentive exists for traders to maintain a peg.
Let us quickly review a simplified version of the rebasing mechanics of vUSD and vETH. For those unfamiliar with the concept, the process is straightforward. At a high level, if the price of vUSD exceeds $1.10 (based on some oracle price feed), then the supply of vUSD expands and all holders receive the same increase of vUSD. Holders include automated market maker smart contracts and other bonding curves. Since the supply in the reserve pool of an AMM for vUSD has increased, the calculation of the price automatically reflects the rebasing event at the specified rate. Similarly, if the price of vUSD is below .90, then the supply shrinks and all the AMM calculated prices are adjusted accordingly.
This process also works in the same way for vETH, i.e. if the price of vETH is greater than 1.10 ETH, then the supply increases and if the price of vETH is less than .90 ETH, then the supply decreases.
Changing the Dynamics of Elastic Stable Assets via Lending Marketplaces
The introduction of a lending marketplace disrupts the purely speculative actions of traders and causes the convergence to a stable market capitalization faster by creating real demand for the pegged asset.
The utility of both vUSD and vETH is quite simple in this model as they are the only assets that may be borrowed. Those looking to borrow vUSD or vETH use approved assets as collateral and take out a secured loan on the vUSD or vETH. Borrowers repay lenders the same amount of vUSD and vETH borrowed plus interest as determined by a bonding curve. Since the supply of vUSD and vETH is elastic, borrowers will have to buy vUSD and vETH on the open market to pay off their loan. In doing so, the interest of the loan creates demand on the vUSD and vETH markets. In essence, lenders of vUSD and vETH are short. On the other hand, those who are borrowing vUSD or vETH are long these assets in the simplest of trading strategies.
More complicated trading strategies could exist for borrowers and a borrower could even be short vUSD or vETH. Consider the following scenario. A borrower Bob is long BTC and puts his liquidity pool tokens of wETH / USDC down as collateral and borrows vUSD. Bob then sells vUSD for BTC and waits for his target BTC price to be reached relative to USD. Since Bob must repay his loan back in vUSD and vUSD will rebase when the price is too far from USD, then this is mostly the same mechanics as if Bob were to wait for his target price of BTC / vUSD to be hit.
Due to the nature of the elastic supply of vUSD and vETH, someone seeking to borrow may be overcollateralized and wish to obtain a loan of vUSD or vETH at an amount that exceeds what is available for lending. Such loan requests could be viewed by market participants as increased demand for the asset the loan is requested to be in. Unfortunately, such requests could be artificial and not indicative of real demand if the potential borrower closes out the loan request prior to it being filled. In order to mitigate such an event, we will start with a simpler model for the lending protocol that does not allow open loan requests. Any improved features will be researched and tested very carefully prior to integration.
Version 1
In the first version of the lending protocol, we will only allow for borrowers to take out a loan based on the current supply and not allow borrowers to make large loan requests (viewed as requests on a bulletin board) that exceed the available supply that can be lent. Such a simplistic model will allow us to observe the market mechanics and give us time to properly conduct research on extensions of this model.
Initially, the protocol will have LP tokens from Uniswap, SushiSwap, Balancer and ValueLiquid as assets that can be used for collateral. The approved liquidity pools are
- wETH-USDC,
- wETH-wBTC,
- wETH-DAI,
- wETH-VALUE and
- wETH-LINK.
In addition, we will integrate Chainlink’s decentralized Price Feeds to verify the value of the collateral of the liquidity pool tokens, as well as the price of vUSD and vETH.
We are exploring two options with Chainlink’s oracle solution to ensure the security of our protocol for the safety of our users.
First, we can use the individual price feeds for ETH, USDC, wBTC, VALUE, and LINK and then calculate the value of the collateralized LP token on the amount of redeemed tokens an LPT would receive.
In the case a liquidity pool’s reserve amount has been manipulated, we would still need to use Chainlink and a solution similar to what we used with our Composite Vaults to calculate the value of the LP token.
Finally, we are exploring Chainlink for providing tamper-proof, up-to-date price data around vUSD and vETH as well.
Future Versions
Non-Value Liquid Pool Tokens as collateralized assets
In subsequent releases, we will allow support for liquidity pool tokens from Uniswap, Sushiswap, and Balancer as collateral. Initially, only the top-weighted liquidity pool tokens will be used but more can be added in the future, particularly by governance proposals by the community. We would most likely impose a higher interest rate for non-Value Liquid Pool Tokens. This difference between the higher interest rate on non-Value Liquid Pool Tokens and VLPTs would then be used to buy Value, vETH, or vUSD and split the profits between Governance Vault Stakers and liquidity pool providers.
Extended Lending Protocol
In future versions, we are considering the following different extensions of our unique lending protocol:
- Allow loan requests to exceed the current supply and perform rebasing events based on the demand for the loans themselves.
- Allow loan requests to exceed the current supply and let market participants satisfy the demand for the loans.
Some features of (2) that would be useful to limit spoofing attacks:
a. fill the loan requests as supply becomes available,
b. allow borrowers to request a minimum loan amount to be satisfied before any borrowing would occur but impose a length of time for the loan to be satisfied before it could be cancelled, and
c. impose a minimum time requirement for a loan request to be outstanding before it can be cancelled and impose a cancellation penalty if it is cancelled prior to that time.
The market mechanics of both are quite complex and intricate and require us to carefully research the various scenarios that can occur under these models. As such, a properly designed extended protocol requires quite a bit more analysis to ensure the protocol is well-functioning and will not negatively impact the Value DeFi community.
Beyond extensions to the protocol, we are also considering an origination fee on loans, similar to how traditional lenders operate. This fee would be used to buy back vUSD or vETH and compensate liquidity providers of the vUSD / USD* and vETH / wETH liquidity pools.
Expectations / Case Studies
We conclude our initial proposal with a few scenarios on how the lending platform could operate.
Early Stages Possibilities
- vUSD Price is $0.30 but a large loan of $2million is requested…
The lending platform launches, people are nervous, not everyone understands it, vUSD is trading at $0.30 and people can’t quite see what will happen.
Lenny comes along, an LP holder and he deposits his $5mil worth of ETH-USDC LP tokens into the Lending platform and requests a $1m loan.
vUSD is only $0.30, and there is only 1million supply.
Over the next few days traders realise this and start to buy vUSD up to $1. Within 10 days vUSD is trading at $2, and so is rebasing daily at 20% daily. The early traders who took the risk and were prepared to buy up vUSD are getting rebased and 10 days later there are 2million vUSDs, trading at $2 each. 2 of the larger buyers/holders of vUSD are now able and willing to provide a loan to Lenny — they deposit 500k vUSD (worth $1m) tokens and Lenny takes the loan at a rate of 20% annualised.
2. Lenny is Happy, Lenders are Happy.
20% annualised for a loan might sound like a lot but Lenny took the loan as he is bullish on the vUSD price. The buyers of vUSD were not so bullish, but they are happy to buy it up cheap and provide the loan and benefit from a few of the rebases and are happy with their 20% annualised roi.
Short vs Long Trader Battles.
- Short trader gets rekt
2 weeks after launch vUSD has grown to be a $20million market cap. 10 million tokens at $2 each. There is $5million worth of LP tokens in the lending platform with $2m of loans lent to them. Lenny thinks the marketcap and token is overvalued and wants to short vUSD. He deposits another $5milion of tokens and requests a further $1m loan of vUSD. He receives the loan and immediately starts to sell it. However 2 other mechanics are against him — 1. Sentiment is generally still bullish for vUSD and people are still wanting to buy more than is available. The price reduces as he sells, he moves the price down to $1.40, but his sells start actively getting bought, and before the end of the day the price is back to $1.80. Lenny was wrong, and now has a tough position — to buy back some/all of his vUSD at a loss before tomorrow’s 8% rebase puts him further underwater… or to extend his short and try again… or to suffer the loss. If he is significantly wrong to short vUSD like this and the price stays at $1.80 for a few more days then this will start to squeeze his position further.
2. Short trader gets liquidated
2 weeks later and the price has stayed at $1.80+ for the whole time, the positive rebase has printed 8% daily to vUSD holders, but not to Lenny as he sold vUSD short. He is under stress as today the price is $2.20 and he sold at $1.40. This alone wouldn’t worry Lenny, he would be happy to wait and look for a cheaper spot to rebuy vUSD.. but over the last 2 weeks the btc AND eth markets have plummeted 30% in value — his LP collateral is only worth$1.4million and he is close to being liquidated. He has a hard decision. He could spend a little bit and buy up $100k worth of vUSD and pay down his debt and keep his collateral comfortably above the liquidation limit.. But Lenny is a gambler and he decides the market will rebound. He goes to sleep. There is a deep crash in the price of ETH and Lenny goes under collateralised on his debt position. He is liquidated 26%.
A portion of his LP position is sold and automatically buys vUSD from ValueDefi and pays his loan back for him, i.e. he isn’t 100% liquidated and his remaining 76% of his LP position tokens are available for him to withdraw (you are never fully liquidated to 0).
3. Short trader makes $15million in 4 days.
2 months later, another short trader, Jenny, has an idea. The vUSD market has been incredibly bullish for this entire time — vUSD is a $450million market cap, the price has touched $3 a couple of times and has been around $1.50 for a week or more. There are 300million vUSD tokens, at a price of $1.50 each. She thinks the market will correct.. She will MAKE it correct. Jenny opens an LP loan request with $50million of LP capital, requesting $20million. In the short term this actually stimulates buy pressure for vUSD and it rises from $1.50 to $2. At this point Jenny’s loan fills and she borrows 10 million vUSD tokens at $2 each — a total of $20million loaned. Jenny immediately starts selling them on the open market. She sells them over 10 hours, and it gradually starts to bring the price down.. Then faster… other traders start to notice, and see that the market is crashing. It causes other traders to sell, to panic, and by the end of the day the price is at $1. The next day there is further doubt and panic in the market as people are realising it might be the end of positive rebases. That night the market is down to $0.80 and the following day there is the first negative rebase of -2%. This further shocks more holders as they see not only their $ value of holdings decrease, but their token amount too. They sell, and within 2 days the price is down to $0.50 and losing 5% a day of their supply in negative rebases. The next morning Jenny buys back the 10 million tokens at $0.50 each, for a total of $5million. She pays back her loan. She has a profit of $15million.
4. Other debtors buy the token back to $1 and pay off their loans.
Shortly after Jenny’s once in a lifetime trade, other people who have taken loans also realise that now would be a good time to buy back the tokens they borrowed cheaply. They start to do so. Some people were simply long on vUSD and held the token and benefited from the positive rebases.. They have had a couple of bad days but are still up overall and pay back the loan with some profit, close their loan positions and just take a few days rest to re-evaluate the market. Others used the loans to trade or act in other markets, they pull money back through and buy back the vUSD — Over the next 2 days $30million further in loans is paid back, with most loan holders having to buy some vUSD back from the open market — there is now buy pressure again and the vUSD price goes from $0.50 and finds it’s peg at $1 again.
The same concept could be applied to vETH (even vBTC, the first rebase token which pegged to BTC for example).
vUSD contract address: https://etherscan.io/token/0x1B8E12F839BD4e73A47adDF76cF7F0097d74c14C
vETH contract address: https://etherscan.io/token/0x76A034e76Aa835363056dd418611E4f81870f16e(edited)
