The admin key risk is real. The article doesn’t discuss that the rising interest rate could attract new lenders to buy out old lenders who want to leave. You don’t need to wait for a borrower to repay the loan you need to wait for anyone to contribute new Dai to the system from anywhere.
It is to my understanding that in order to borrow money through compound you have to “over collateralize” your loan. However, normally people take out loans because they need that money to do something else. Why would someone lock up, let’s say $300, just to borrow $100 with interest? Why not just use the $300 they initially had and not pay any interest? Thanks in advance
This type of DeFi lending is still in its nascent stages. MakerDao started off lending against ETH at 0.5%. At that rate, it makes a lot of sense to avoid tax events and borrow against what you think is an appreciating asset for a variety of reasons not the least of which is to avoid the opportunity cost of selling your ETH too low.
Even at current rates, it still might make sense in some circumstances. I mean people routinely borrow at exorbitant interest rates for cars, and many lower tier buy here pay here auto loans are overcollateralized.
Overcollateralization is quite common against illiquid assets. When DeFi matures and demand stabilizes the rate to be more predictable, there might again be a sound financial reason to collateralize your ETH to borrow for any number of reasons. Right now with ETH's current valuation direction being questionable at bet in the short term, I think the primary reason why people are borrowing against their ETH is to buy more ETH or other crypto to speculate on price.
If you have ETH that’s worth $30,000 and you want to borrow $10,000 without triggering a tax event, that’s a possibility. (Correct me if I’m wrong)
If you wanted to keep your money in ETH and have some USD available, you can use that USD to buy more ETH. So instead of owning 1 ETH, you could in theory own 1.33 ETH.
Obviously, this doesn’t work out well if the price drops.
It’s called ‘leverage’. You borrow to buy more, so when price goes up, your profit has a multiplier.
The fact why you borrow against your own collateral and not just sell, is you want to minimize your opportunity costs. Meaning: if you sell and price goes up, you missed making a profit. With a system like MakerDAO you can keep your speculative asset, borrow against it as collateral and either leverage up (by buying more of the asset) or just using the cash for real world purposes.
It is at best as decentralized as the ethereum network. The admin functions make it less decentralized, but the baseline execution of the smart contracts and interest accrual is decentralized.
Compound's current contracts are decentralized, but not immutable, and therefore somewhat custodial.
Is it though? There's no mention of spank in the article other than to mention how much DAI they were holding, and I appreciate knowing the affiliation of the author before reading it in case there is bias present.
Note: I don't own and am not interested in owning spank, nor do I know much about it.
Thought experiment: with Compounds 98.62% utilization rate, what USD or ETH amount would it take to artificially freeze liquidity, and induce a bank run to drive down the panic sell exchange rate of cDAI and scoop up cheap cDAI to be repaid a short time later at a more nominal value once the dust settles.
After reading the open Zeppelin audit report, there are a number of other decisions made by compound that seem amateurish at best.
For instance, compound currently calculates interest on the assumption that each block is 15 seconds apart instead of just using the timestamps of the blocks. Also, there are rounding errors when calculating the interest that open up the system to some kinds of attacks. I would really like to see a compound v3 before I put my dai in it.
The contracts don't assume that blocks are 15 seconds, its just a heuristic for giving people an idea of the APR off-chain. On the chain, the formulas are exact and in terms of blocks, because that \*is\* time on the EVM. The contracts utilize borrowRatePerBlock and supplyRatePerBlock. If the math had worked in terms of timestamps instead of blocks, the criticism would have been that timestamps can be manipulated by miners. Auditors are paid to find issues - as many as they can - and I personally view it as a credit to the transparency of Compound that the company was so eager to publish the results. OpenZeppelin was actually surprised that these results were to be published immediately, since that's generally not how they work.
The [1inch.exchange](https://1inch.exchange) guys privately [noticed this](https://i.imgur.com/YdUzd2z.png) before the Zeppelin audit report came out when they were comparing Fulcrum's iDAI to Compound's cDAI.
Compound.Finance interest is **compounded** at 15 seconds. *"Interest can be compounded on any given frequency schedule, from continuous to daily to annually."* *
Compound.Finance probably selected the K.I.S.S. principle. APR can be easily computed given the interest **rate** and compound frequency (15 seconds). If they based their compound frequency on the block times, imaging the difficulty for the user to compute the APR.
*Compound interest source:
I think the narrative of Compound (and other DeFi platforms like it) as a "better savings account" is really dangerous. Not mentioned in the article is the additional layers of risk in DAI, interest rate risk, and black swan risk. I worry that the neglect of these risks in the narrative will lead to people losing what they can't afford to lose.
For the time being (and likely the foreseeable future), these platforms are clearly risk assets and should be treated as such (i.e., don't invest what you can't afford to lose, diversify across platforms and across lending assets, and take steps to protect yourself).
All that said, I believe they perform favorably (especially because much of the risk is uncorrelated to the market as a whole and also much of it is not correlated between different platforms) and at current rates, they should likely be part of a diversified portfolio.
Another thing I wish to add is my commentary to the quote below:
The only tool that Compound has at their disposal to address this is to use the centralized administrator to upgrade their interest rate model, which is exactly what they did 6 weeks ago when the utilization rate increased to ~99% (same time as the quoted tweet above).
Adjusting the interest rate only works during a period of price stability or bear market. Otherwise, in a persistent bull market (which is a matter of when, not if), raising the interest rate will not matter at all. Thus in times of "crisis", raising the interest rate is actually a non-solution.
> Adjusting the interest rate only works during a period of price stability or bear market. Otherwise, in a persistent bull market (which is a matter of when, not if), raising the interest rate will not matter at all. Thus in times of "crisis", raising the interest rate is actually a non-solution.
No need to speak in absolutes. It'll absolutely make a difference, we can see it in exchange behavior - exchanges with low lending fees have much more volume than exchanges with high lending fees, contributing factors notwithstanding.
It will matter... just not enough. And frankly, what did you expect? You loan you asset out, you're at the mercy(ish) of the borrower. Duh?
I break my investigation down by category below, but the most important things to know are:
The smart contract security seems legit.
Compound is a CUSTODIAL system, all lending pools can be trivially drained if their admin private key is compromised.
When you lend on Compound, you are NOT guaranteed to be able to withdraw whenever you want. If you try to withdraw your funds and all the money is locked up in outstanding loans, your withdrawal transaction will fail.
It would take administrative action to change the rate. Right now the article states,
> For cDAI, the Base Rate = 5% and the Multiplier = 15% (the values are hardcoded into the contract). At a 100% utilization rate the interest paid by borrowers would be 20%.