By George Liu and Matthew Turk
Tl;dr: This weblog analyzes centralized stablecoin lending yield for Compound Finance and shares our insights on efficiency, volatility, and elements that drive this yield on collateralized lending of stablecoins in DeFi. The evaluation reveals that this lending yield can outperform the risk-free yield within the TradFi market.
Partially two of this quantitative analysis piece, we are going to study stablecoin lending yield for the Compound Finance V2 decentralized finance (DeFi) protocol and share our insights on yield efficiency, volatility, and what elements are driving yield on collateralized lending of stablecoins by way of DeFi protocols. We additionally evaluate the “risk-free” yield in conventional finance (TradFi) to the idea of “low-risk” yield in DeFi, which we launched partially one.
ACKNOWLEDGEMENT: Whereas we’re conscious of the latest collapse of Terra’s algorithmic stablecoin TerraUSD (UST), our evaluation right here is on the world of collateralized lending yield for centralized stablecoins. We’re centered particularly on Compound for USDC and USDT (fiat-backed stablecoins), which have disparate dangers and alternatives.
We conclude on this piece that utilizing stablecoins for low-risk (inside DeFi) collateralized lending might outperform the risk-free funding within the conventional monetary market.
As talked about partially one in all this weblog put up, a Compound person who has positioned their property right into a liquidity pool can calculate whole lending yield utilizing exchangeRate, which is a sign of the worth of the curiosity that the lender can anticipate to obtain over time, and the return from time T1 to T2 might be merely obtained as
R(T1,T2)=exchangeRate(T2)/exchangeRate(T1)-1.
Moreover, annualized yield for this sort of collateralized lending (assuming steady compounding) might be calculated as
Y(T1,T2)=log(exchangeRate(T2)) — log(exchangeRate(T1))/(T2-T1)
Whereas the Compound liquidity swimming pools help many stablecoin property such USDT, USDC, DAI, FEI and so forth, we’re solely going to investigate the highest 2 stablecoins right here, i.e USDT and USDC, which have a market capitalization of $80B and $53B respectively. Collectively, they make over 70% of the whole market of the stablecoins.
Under are the plots of the annualized each day, weekly, month-to-month and biannual yields generated in accordance with the formulation within the earlier part. The each day yield is considerably risky, whereas the weekly, month-to-month and biannual yields are respectively the smoothed model of the prior granular plot. USDT and USDC have comparatively related patterns within the plot, as they each expertise excessive yield and excessive volatility in the course of the begin of 2021. This means that there are some systematic elements which might be affecting the stablecoin lending market as a complete.
Supply: The Graph
One speculation of the systematic elements that might have an effect on the lending yield are crypto market knowledge (just like the BTC/ETH costs) and its corresponding volatilities. When BTC and ETH are in an ascending development, some bull-chasing traders might borrow from the stablecoin swimming pools to purchase BTC/ETH, after which use the bought BTC/ETH as collateral to borrow extra stablecoins and repeat this cycle till their leverage reaches the specified stage. Moreover, when the market enters right into a excessive volatility regime, there are extra centralized and decentralized crypto transactions which might enhance the demand for stablecoins as properly.
Now, to verify the connection of the stablecoin yield and the crypto market knowledge, we carry out a easy linear regression evaluation to see how a lot variation within the yield might be attributed to the value and volatility elements utilizing the next formulation:
To measure the magnitude of those elements’ contribution, we use the R-Squared rating, which has a variety of [0, 100%]. A rating of 100% would imply that the yield is totally decided by the contributing elements.
Regression of USDC/USDT on the BTC market and the ETH market respectively lead us to the next R-Squared desk:
ETH market knowledge has a greater explanatory energy (18% & 17%) than the BTC market knowledge (16% & 11%) in figuring out the yield of USDC and USDT. That is unsurprising, significantly attributable to ETH’s elevated reputation and expanded footprint within the DeFi market because the begin of 2021. As seen with these outcomes, crypto value and volatility elements didn’t totally clarify the yield in stablecoins. We will conclude that there have to be different elements that assist to enhance the rating from the fundamental mannequin.
We carried out additional exploratory evaluation by introducing the historic stablecoin provide knowledge and MACD technical indicator value knowledge to the mannequin. The stablecoin provide (the whole variety of stablecoins provided to Compound liquidity swimming pools) ought to — intuitively — have an effect on the provision/shortage of the stablecoins and not directly affect the yield. MACD is a vital momentum buying and selling sign (subtracting the 26 interval EMA from the 12 interval EMA — on this case on value) because it might assist momentum traders to resolve when to leverage and when to deleverage.
We see a noticeable enhance in R-Squared scores, as each USDC and USDT bought a bump to a stage round 60%-70% as proven beneath.
From this knowledge we are able to conclude that stablecoin provide is a considerable contributing issue, because it alone is ready to convey the rating to round 60% for each stablecoins in any of the 2 markets. It appears to counsel that [supply] is a significant factor in affecting the yield within the stablecoin lending market. That is similar to the TradFi world, the place credit score provide by the Federal Reserve will have an effect on the overall rate of interest of the entire system.
The introduction of MACD knowledge (on BTC and ETH value) brings blended enchancment. Within the case of the BTC market, its impartial contribution is way lower than the availability issue, and the marginal profit over the shoulder of provide is just a few share factors. We observed within the ETH market, nevertheless, that MACD has a better impartial contribution to the R-Squared worth as in comparison with the BTC market. This implies that stablecoin lending yields are extra correlated with momentum based mostly buying and selling exercise in ETH than in BTC.
An instance of the regression coefficients for USDC lending yield within the ETH market are displayed beneath. The desk means that larger ETH costs, volatility and [stable coin supply] are usually related to decrease USDC lending yield. On the similar time, the stronger the MACD sign is, the upper the yield would go.
Whereas it’s attention-grabbing to disclose what has pushed the low-risk yield on stablecoin lending, it is usually vital to match these yields with the counterpart within the TradFi market.
As a result of stablecoin lending yields are derived from the realized floating rates of interest for collateralized loans on the Compound platform, we chosen the Common Collateral (GC) charge used within the conventional cash market because the comparable risk-free charge, as a result of it is usually a floating charge with treasury debt because the mortgage collateral.
Under is a plot of the portfolio worth of the investments that earn USDC lending yield, USDT lending yield, and GC charge yield respectively. The investments all begin with $100 preliminary worth on 2020–05–01, and finish on 2022–05–01. As seen beneath, yield on USDT and USDC collateralized lending is larger than the GC charge by a big margin. However, risk-free funding that earns GC charge hardly grows for a similar interval.
The typical rate of interest within the desk beneath additionally confirms that GC charge is on common round 0.08%, whereas USDC and USDT lending yields are respectively 3.71% and 4.51% for this era as seen beneath. (We additionally checked the 2Y time period yield on the treasury debt on 2020–05–1 which is merely 0.2%)
For the foreseeable future, it’s affordable to conclude that the low-risk charge, inside the crypto market not less than, will proceed to outperform the risk-free charge within the TradFi market. One cause for that is the sensible contract danger, or liquidation danger talked about partially one in all this weblog. Nevertheless, a bigger cause is the slower progress within the stablecoin provide relative to the expansion within the crypto economic system as a complete. By comparability, the TradFi market has seen main credit score progress because the begin of the Covid-19 pandemic, which has helped to drive the risk-free charge to historic lows (see Fed steadiness sheet progress beneath).
This weblog offered a broadly indicative evaluation of the low-risk yields obtainable from collateralized lending of stablecoins by way of DeFi protocols. Whereas these yields could also be very risky each day, their common development might be defined comparatively properly by BTC/ETH costs, volatilities, stablecoin provide and MACD (momentum buying and selling actions). We additionally in contrast these yields with the risk-free charge within the TradFi market the place we see constant outperformance within the crypto market. To reiterate, this isn’t monetary recommendation.
Subsequent steps
We, as a part of the Knowledge Science Quantitative Analysis staff, purpose to get a holistic understanding of this house from a quantitative perspective. We’re searching for individuals which might be passionate on this effort to hitch our rising staff. In case you are thinking about Knowledge Science and particularly Quantitative Analysis in crypto, come be part of us.
The evaluation makes use of the Compound v2 subgraph made obtainable by way of the Graph Protocol. Particular because of Institutional Analysis Specialist, David Duong, for his contribution and suggestions.