There are two ways to write PSM Risk Assessment. One that is isolated from current 101% LR stablecoins implementation and one that compares PSM with current implementation of low LR stablecoin vaults and its impact on risk. I decided to choose the latter since risks with 101% LR stablecoin vaults were already discussed a couple of times and are well known to the community. Yet, PSM does increase stablecoin exposure among other things and both risks and benefits need to be properly addressed.
In this risk assessment I want to primarily focus on the following topics:
- Impact on increased stablecoins exposure
- Risk compensation of increased stablecoin exposure
- Impact on market making activities
- Additional features of PSM
Increased stablecoin exposure
MKR holders have voted yes (with only 55% support though) to support PSM implementation with 0.1%
tin and 0.1%
tin parameter of 0.1% effectively means that governance is capping DAI price to $1.001 (assuming stablecoins in PSM are worth $1). The most liquid venue for stablecoin trading is currently Curve where DAI price in large extent gets defined. Looking at the bid side of Curve’s orderbook, we can see that 93m DAI needs to be sold into Curve pools to reach the DAI/USDC price of 1.001. If we look at other important stablecoins pairs such as ones with USDT and TUSD, we can estimate additional 90m DAI selling is needed, bringing the aggregate amount to around 180m DAI. Finally, there are other venues where DAI is in demand over $1.001 price mark and we could easily estimate that additional 250m DAI minted from stablecoins might be needed to reduce DAI price below $1.001.
This means Maker would increase DAI backing with stablecoins from current 45% towards 55% if a
tin parameter of 0.1% is selected. Note that this is only estimation and situation can change until PSM implementation is effective (such as recent proposal of Compound to lower DAI borrowing on their platform which might put pressure of the peg)
DAI/USDC orderbook on Curve
By implementing PSM, risk profile of Maker should worsen since additional stablecoins are onboarded and low fee collection will probably not offset increased risks. This is because only 0.1% fee is collected on stablecoin exposure versus 4% in the past (actually only 1% if assuming recent 101% LR). However the 0.1% is one time fee collection “on the way in’’ and can multiply on a yearly basis if DAI demand and supply imbalances cause DAI to hit $0.999 floor and $1.001 ceiling frequently. It is questionable though if a high turnover can be reached to accumulate yearly fees well over 1%, yet it is theoretically possible, especially if Maker starts competing with other stablecoin trading venues (addressed later).
However I want to note that PSM is much more efficient, when it comes to fee collection versus current 101% LR stablecoin vault implementation. The maximum amount of fees Maker could collect with 101% LR stablecoin vault was 1%, which may sound better than 0.1% (even though one time), but there were many known issues with unwinding of such vaults and actual collection of those fees.
Impact on market making ecosystem
Community discussed the impact of PSM on market making (MM) ecosystem and primarily exposed the problem of MM losing incentive or opportunity to arbitrage the peg the way they were used to until now. In the pre-low LR stablecoin vaults world, speculators could speculate on the price and arb it very freely. Once 101% LR stablecoin vaults were established, this arbitrage was reduced severely. If and when PSM with 0.1%
tin is implemented, the arbitrage MM used to do is more or less gone.
However there is a different type of arbitrage now. It isn’t as speculative anymore, but has direct realization of arbitrage profits. Every time DAI price leaves the
tout bounds, everybody can immediately realize profit by trading through PSM. This is much different from how DAI peg relied on MM in the past, where MM needed some assurance governance will rebalance demand and supply so that DAI reaches $1 peg again and they could exit their positions and realize profits. In that sense, the arbitrage opportunity under PSM is still present, but 1) is way less speculative than it was in the past, 2) doesn’t enable high returns as in the past but on the other hand 3) enables direct realization of arbitrage profits.
One feature that PSM enables for DeFi ecosystem is zero slippage stablecoin trading. However it depends on debt ceilings and type of stablecoin that will be supported. Maker will therefore directly compete with AMMs like Curve, although it would still charge higher fees (0.1%
tout versus 0.04% Curve fee). Yet it will offer zero slippage and traders performing larger trades might want to utilize Maker instead of other AMMs.
Another feature of PSM is that it will enable Maker governance to follow DAI demand and supply imbalance. With a more narrowly defined DAI price, each demand or supply shock would be directly observed by the amount of stablecoin collateral held in the PSM.
Finally, PSM enables easier unwinding of current low LR stablecoins vaults. Vault owners might be more tempted to exit their positions when DAI price gets decreased due to PSM and governance would have less issues with coordinating liquidation of such vaults. There were also ideas that these vaults could be directly liquidated to PSM using specific techniques.
To conclude, PSM with proposed
tout parameters offers an alternative way of peg management for Maker. It starts relying more on stablecoins and their pricing, arbitrage becomes different and risk compensation gets reduced. However, PSM is much more effective than low LR stablecoin implementation.
@Risk understands that the governance is in favour of lowering the business risks by having DAI price much closer to $1 and less volatile and is prepared to sacrifice increased collateral risk profile (more stablecoin exposure with much lower risk compensation). As I have mentioned in the past, PSM is really focused on DAI peg, but worsens collateral risk. So it becomes a question of business risk versus collateral risk to which this assessment doesn’t provide an answer. Negative rates are potentially a better implementation in terms of collateral risk, but possibly much worse from a business risk perspective and peg management efficiency. Nevertheless I think everybody agrees that the primary goal still stands to onboard RWA and other yielding collateral (such as CropJoin implementation) that scales DAI supply and reduces PSM’s risk profile.