9 ms·

A lot of Ethereum DeFi (and cryptocurrency in general) has been pretty frustrating to watch, because it's a lot of people with big ideas and little understandin

by Taek 7y ago

A lot of Ethereum DeFi (and cryptocurrency in general) has been pretty frustrating to watch, because it's a lot of people with big ideas and little understanding of how to build stable financial systems. In the case of Maker, what's interesting is that a stablecoin is actually possible. Maker has a really good core idea, several great elements to it: 1. The stablecoin is backed by collateral, typically a significant amount more collateral than the amount of stablecoin that has been issued. 2. People locking up collateral get greater exposure to Eth's volatility. So there's genuinely a reason to lock up collateral to issue stablecoins if you are bullish on Eth. But, this breaks down a bit. There's no way to instantly convert the stablecoin back into the collateral asset, you have to find someone with a CDP who is willing to buy the stablecoin from you. Similarly, there's no way to unlock your collateralized asset, you have to find someone who is willing to sell you the stablecoin in order to open up your vault and get your eth out. So on both sides of the equation, you have this liquidity risk that doesn't really need to exist. I could go on for a bit more, there are a bunch of other design decisions that are backwards and don't work well. But the fundamental ideas are actually pretty okay, and if you took a more traditional finance person and had them work through all the tiny details, I think you'd end up with something pretty powerful. In that sense, it's almost like encryption. The tiniest detail being incorrect can make the whole construction useless, even if as a whole the fundamentals are pretty solid.

enjeyw 7y ago
>So on both sides of the equation, you have this liquidity risk that doesn't really need to exist. I'd love to hear more of your thoughts on this. It's apparently big problem with the Maker protocol, and seems to be an inherent issue with the use of a perpetual (afaik we're seeing the price of DAI spike as a consequence), but I'm not aware of a simple fix.
- Taek 7y ago
  A good fix requires the whole system to be re-imagined a bit. You want to start with the invariant that a stablecoin holder can liquidate to Eth at any point, and that a collateral issuer can extract their Eth at any point. Instead of having the collateral provider be the only one who can issue the stablecoin, you could build a system where anyone could acquire the stablecoin (Dai) by putting in an appropriate amount of Eth. Any extra Eth that gets put into the pool will issue a volatile asset, which I'll call Vai here for simplicity. Where Dai has no exposure to volatility, Vai has increased exposure to volatility. As the Eth price rises, the amount of Eth that Vai can be redeemed for increases, and as the Eth price drops, the amount of Eth that Vai can be redeemed for decreases (this is the opposite of Dai). At any time, Eth and Vai can be redeemed for Eth from the pool. As the price of Eth moves around, the pool becomes over-collateralized and under-collateralized. If at any point the pool becomes under-collateralized, all Vai becomes worth zero, and all Dai can redeem Eth proportional to the amount of collateral that there is. Meaning, in a "black swan" scenario, Dai holders become exposed at 1x to the price volatility of Eth. This is much better than the whole system melting down, and the Dai holders do get to dodge a significant amount of the downtrend while the Vai evaporates. (Dai holders see no exposure until all the Vai is gone). You can keep the pool balanced using interest rates. If you have a target collateralization level (say 150%), then any time the pool is under-collateralized, you can automatically drop the interest rate that Dai holders earn (going negative if necessary). If the pool is over-collateralized, you increase the interest rate. A lower interest rate encourages Dai holders to withdraw and Vai holders to enter. A higher interest rate encourages Dai holders to enter and Vai holders to withdraw. ----- The most important thing about the system above is that it is highly predictable. In all scenarios where the price of Eth moves, you can model what happens, and you have guarantees on how much Eth you can draw, and you can withdraw that Eth immediately (well, you'll need to pay blockchain fees and wait for a block, but you don't need a counterparty or any sort of system liquidity). The worst case for the stablecoin holders is that they get exposure to 1x Eth, but this only happens if Eth drops faster than people deposit Vai. There are a bunch of math tricks you can use to allow people to select different exposure levels (you could have some Dai targeting 250% over-collateralization, meaning it's very robust to huge drops in Eth price, and other Dai only target 15% over-collateralization, and these two assets could be fully fungible against eachother) for Dai holders and Vai holders, but we are getting beyond the scope of a single HN comment.
lebed2045 7y ago
this was my exact thoughts at the end of 2017 when Maker released their first version. Why instead of CDPs just don't use one big liquidy pool which initially funded with sufficient resources? If it is big enough - it would initially overcollateralized the system (when the cap is small) and can be used as a stabilization reserve, i.e. everyone would able to redeem their $1 stablecoin for $1 of ETH at any moment of time. I even try to design such a system https://stableunit.org/StableUnit-whitepaper.pdf https://stableunit.org/StableUnit-whitepaper.pdf The main assumption that stablecoin with collateral/reserve has two values: 1) "redeemability value", i.e. the token without demand is worth $1 because you can redeem it for $1 of another asset with demand such as ETH. 2) network/native value: more people use the coin, higher the value. In the same manner as Dogecoin worth anything at all, like Metcalfe's law but for cryptocurrency. With an empirical estimation of growth asymptotic higher than NlogN. The important part here that as the system grows bigger 1) grows linear but 2) grows to faster than linear which means there are some sizes of then system when 2) > 1). This practically means that you can back part of stablecoin with less liquid assets because they won't be used often (if all). It's one way to explain why USDT can function even with partially collateralized. The problem with all of this, that one thing is to have system design (I'm convinced that there are many different decentralized stablecoin systems that can work) and another to build a working project with user adoption. I've spoken with many people at MakerDAO and they all know about the limitations/imperfection of the system. For oracles, they chose that expensive design because it was an "engineering approach which just works security". And as we see right now at stablecoins market cap, they were right to do so.
daveytea 7y ago
> There's no way to instantly convert the stablecoin back into the collateral asset, you have to find someone with a CDP who is willing to buy the stablecoin from you. This is incorrect and actually doesn't make sense. An important part of DeFi are DEXes (decentralised exchanges) such as Uniswap and Kyber. There is a liquidity pool where you can instantly buy/sell assets. There is no 'waiting' to find someone (i.e. no problems with coincidence of wants). You definitely do not need to find someone with a CDP. You can hold DAI without opening a CDP. Contracts can hold DAI (as DEXes do). > Similarly, there's no way to unlock your collateralized asset, you have to find someone who is willing to sell you the stablecoin in order to open up your vault and get your eth out. Also incorrect. You need to pay back the debt of your CDP with DAI. You can do this easily by buying DAI on a DEX, paying back the debt, then releasing your collateral. Some services exist to do this in 1 transaction, so you don't need to actually 'buy' any other asset. You just send the transaction to a contract and they take care of the details. The DeFi, specifically the Ethereum space, has moved very quickly in a short amount of time, so there are a lot of new concepts and instruments out there. I think a more traditional finance person will have trouble understanding it all as in some cases, there are no analogies to the traditional finance system (e.g. flashloans).
- wayoutthere 7y ago
  > I think a more traditional finance person will have trouble understanding it all as in some cases, there are no analogies to the traditional finance system (e.g. flashloans). All "cryptocurrencies" are, in effect, complex securities that are derivatives of greater market factors. The crisis of 2007 showed the risk of using complex investment vehicles that were poorly understood by investors -- and the CDOs that were sold in the mid 00s were far more transparent and predictable than cryptocurrency. Given that crypto will always be one asset class among many, if the traditional finance system (which already operates outside any single fiat currency) can't understand it, they won't use it for anything more than speculation. Using crypto as an investment vehicle requires a reasonably accurate assessment of risk. In the case of fiat currencies, central banks manage that risk so that investors can rely on the liquidity of the overall system without wiping out deposits. DeFi has no such mechanism, and no central bank to absorb a big hit temporarily in the case of a black swan event. I still feel that cryptocurrency is just the 21st century version of penny stocks and junk bonds. Fiat currency works because its power as currency is secured by a government able to mobilize military and industrial power to solve market problems. The most heavily traded currencies (RMB, USD, EUR) are those backed by large industrial and military powers because those countries have the scale and political power to manage market risk. Centralized governance is a feature, not a bug.

by Taek 7y ago

enjeyw 7y ago

>So on both sides of the equation, you have this liquidity risk that doesn't really need to exist. I'd love to hear more of your thoughts on this. It's apparently big problem with the Maker protocol, and seems to be an inherent issue with the use of a perpetual (afaik we're seeing the price of DAI spike as a consequence), but I'm not aware of a simple fix.

Taek 7y ago
A good fix requires the whole system to be re-imagined a bit. You want to start with the invariant that a stablecoin holder can liquidate to Eth at any point, and that a collateral issuer can extract their Eth at any point. Instead of having the collateral provider be the only one who can issue the stablecoin, you could build a system where anyone could acquire the stablecoin (Dai) by putting in an appropriate amount of Eth. Any extra Eth that gets put into the pool will issue a volatile asset, which I'll call Vai here for simplicity. Where Dai has no exposure to volatility, Vai has increased exposure to volatility. As the Eth price rises, the amount of Eth that Vai can be redeemed for increases, and as the Eth price drops, the amount of Eth that Vai can be redeemed for decreases (this is the opposite of Dai). At any time, Eth and Vai can be redeemed for Eth from the pool. As the price of Eth moves around, the pool becomes over-collateralized and under-collateralized. If at any point the pool becomes under-collateralized, all Vai becomes worth zero, and all Dai can redeem Eth proportional to the amount of collateral that there is. Meaning, in a "black swan" scenario, Dai holders become exposed at 1x to the price volatility of Eth. This is much better than the whole system melting down, and the Dai holders do get to dodge a significant amount of the downtrend while the Vai evaporates. (Dai holders see no exposure until all the Vai is gone). You can keep the pool balanced using interest rates. If you have a target collateralization level (say 150%), then any time the pool is under-collateralized, you can automatically drop the interest rate that Dai holders earn (going negative if necessary). If the pool is over-collateralized, you increase the interest rate. A lower interest rate encourages Dai holders to withdraw and Vai holders to enter. A higher interest rate encourages Dai holders to enter and Vai holders to withdraw. ----- The most important thing about the system above is that it is highly predictable. In all scenarios where the price of Eth moves, you can model what happens, and you have guarantees on how much Eth you can draw, and you can withdraw that Eth immediately (well, you'll need to pay blockchain fees and wait for a block, but you don't need a counterparty or any sort of system liquidity). The worst case for the stablecoin holders is that they get exposure to 1x Eth, but this only happens if Eth drops faster than people deposit Vai. There are a bunch of math tricks you can use to allow people to select different exposure levels (you could have some Dai targeting 250% over-collateralization, meaning it's very robust to huge drops in Eth price, and other Dai only target 15% over-collateralization, and these two assets could be fully fungible against eachother) for Dai holders and Vai holders, but we are getting beyond the scope of a single HN comment.

lebed2045 7y ago

this was my exact thoughts at the end of 2017 when Maker released their first version. Why instead of CDPs just don't use one big liquidy pool which initially funded with sufficient resources? If it is big enough - it would initially overcollateralized the system (when the cap is small) and can be used as a stabilization reserve, i.e. everyone would able to redeem their $1 stablecoin for $1 of ETH at any moment of time. I even try to design such a system https://stableunit.org/StableUnit-whitepaper.pdf https://stableunit.org/StableUnit-whitepaper.pdf The main assumption that stablecoin with collateral/reserve has two values: 1) "redeemability value", i.e. the token without demand is worth $1 because you can redeem it for $1 of another asset with demand such as ETH. 2) network/native value: more people use the coin, higher the value. In the same manner as Dogecoin worth anything at all, like Metcalfe's law but for cryptocurrency. With an empirical estimation of growth asymptotic higher than NlogN. The important part here that as the system grows bigger 1) grows linear but 2) grows to faster than linear which means there are some sizes of then system when 2) > 1). This practically means that you can back part of stablecoin with less liquid assets because they won't be used often (if all). It's one way to explain why USDT can function even with partially collateralized. The problem with all of this, that one thing is to have system design (I'm convinced that there are many different decentralized stablecoin systems that can work) and another to build a working project with user adoption. I've spoken with many people at MakerDAO and they all know about the limitations/imperfection of the system. For oracles, they chose that expensive design because it was an "engineering approach which just works security". And as we see right now at stablecoins market cap, they were right to do so.

daveytea 7y ago

> There's no way to instantly convert the stablecoin back into the collateral asset, you have to find someone with a CDP who is willing to buy the stablecoin from you. This is incorrect and actually doesn't make sense. An important part of DeFi are DEXes (decentralised exchanges) such as Uniswap and Kyber. There is a liquidity pool where you can instantly buy/sell assets. There is no 'waiting' to find someone (i.e. no problems with coincidence of wants). You definitely do not need to find someone with a CDP. You can hold DAI without opening a CDP. Contracts can hold DAI (as DEXes do). > Similarly, there's no way to unlock your collateralized asset, you have to find someone who is willing to sell you the stablecoin in order to open up your vault and get your eth out. Also incorrect. You need to pay back the debt of your CDP with DAI. You can do this easily by buying DAI on a DEX, paying back the debt, then releasing your collateral. Some services exist to do this in 1 transaction, so you don't need to actually 'buy' any other asset. You just send the transaction to a contract and they take care of the details. The DeFi, specifically the Ethereum space, has moved very quickly in a short amount of time, so there are a lot of new concepts and instruments out there. I think a more traditional finance person will have trouble understanding it all as in some cases, there are no analogies to the traditional finance system (e.g. flashloans).

wayoutthere 7y ago
> I think a more traditional finance person will have trouble understanding it all as in some cases, there are no analogies to the traditional finance system (e.g. flashloans). All "cryptocurrencies" are, in effect, complex securities that are derivatives of greater market factors. The crisis of 2007 showed the risk of using complex investment vehicles that were poorly understood by investors -- and the CDOs that were sold in the mid 00s were far more transparent and predictable than cryptocurrency. Given that crypto will always be one asset class among many, if the traditional finance system (which already operates outside any single fiat currency) can't understand it, they won't use it for anything more than speculation. Using crypto as an investment vehicle requires a reasonably accurate assessment of risk. In the case of fiat currencies, central banks manage that risk so that investors can rely on the liquidity of the overall system without wiping out deposits. DeFi has no such mechanism, and no central bank to absorb a big hit temporarily in the case of a black swan event. I still feel that cryptocurrency is just the 21st century version of penny stocks and junk bonds. Fiat currency works because its power as currency is secured by a government able to mobilize military and industrial power to solve market problems. The most heavily traded currencies (RMB, USD, EUR) are those backed by large industrial and military powers because those countries have the scale and political power to manage market risk. Centralized governance is a feature, not a bug.