Original source: Coinbase
Original compilation: ChinaDeFi
Original compilation: ChinaDeFi
In this quantitative research article, we’ll look at the Compound Finance V2 DeFi protocol’s stablecoin loan yield and share our thoughts on yield performance, volatility, and what drives DeFi protocol collateralized lending yields.
While we are aware of the recent debacle of Terra’s algorithmic stablecoin, TerraUSD (UST), our analysis here is about the collateralized lending yield space of centralized stablecoins.
We conclude in this post that low-risk (in the DeFi context) mortgages using stablecoins can outperform risk-free investments in traditional financial markets.
USDT/USDC Yield Analysis
Compound users who have placed assets into a liquidity pool can use exchangeRate to calculate the total lending yield, which indicates the value of interest that lenders can earn over time, available from time T1 to T2:
R(T1,T2)=exchangeRate(T2)/exchangeRate(T1)-1
Additionally, the annualized rate of return (assuming continuous compounding) for this type of mortgage can be calculated as:
Y(T1,T2)=log(exchangeRate(T2)) — log(exchangeRate(T1))/(T2-T1)
Although the Compound liquidity pool supports USDT, USDC, DAI, FEI and other stablecoin assets, here we only analyze the first two stablecoins, namely USDT and USDC, whose market capitalizations are $80 billion and $53 billion respectively. Together, they account for more than 70% of the total stablecoin market.
image description
Source: the graph
Systemic factors that may affect loan yields are crypto market data (such as the price of BTC/ETH) and their corresponding volatility. When BTC and ETH are on an upward trend, some bull-chasing investors may borrow money from the stablecoin pool to buy BTC/ETH, then use the purchased BTC/ETH as collateral to borrow more stablecoins, and repeat the cycle. until the leverage reaches the desired level. Additionally, when the market enters a high volatility regime, there will be more centralized and decentralized crypto exchanges, which will also increase the demand for stablecoins.
Now, to examine the relationship between stablecoin returns and cryptocurrency market data, we perform a simple linear regression analysis using the following formula to see how much of the change in returns can be attributed to price and volatility factors:
To measure the influence of these factors, we use the R-Squared score, which ranges from [0,100%]. A score of 100% means that the rate of return is entirely determined by influencing factors.
We regress USDC/USDT in the BTC market and the ETH market respectively, and get the following R-Squared table:
ETH market data (18% and 17%) have better explanatory power than BTC market data (16% and 11%) in determining the yield of USDC and USDT. This is not surprising, especially due to the growing popularity and breadth of ETH in the DeFi market since the beginning of 2021. As can be seen from these results, cryptocurrency price and volatility factors do not fully explain stablecoin returns. We can conclude that there must be other factors that contribute to improving the base model's score.
We conducted a further exploratory analysis of the model by introducing the historical supply data of stablecoins and the MACD technical indicator price data. Stablecoin supply (the total number of stablecoins provided to Compound's liquidity pool) should intuitively affect stablecoin availability/scarcity and indirectly affect yield. The MACD is an important momentum trading signal because it helps investors decide when to leverage and when to deleverage.
We saw a significant improvement in the R-Squared score, with both USDC and USDT reaching a level of around 60%-70%, as shown in the chart below.
From this data we can conclude that the supply of stablecoins is an important factor as it enables the scores of stablecoins in either market to be around 60%. This seems to indicate that supply is a major factor affecting yields in the stablecoin lending market. This is very similar to the traditional economic world.
The release of MACD data (regarding BTC and ETH prices) brought mixed improvements. Taking the BTC market as an example, its independent contribution is far less than the supply factor, only a few percentage points beyond the marginal benefit of supply. However, we note that MACD has a larger independent contribution to R-Squared in the ETH market compared to the BTC market. This suggests that stablecoin lending yields are more correlated with momentum-based trading activity in ETH than in BTC.
The figure below is an example of the regression coefficient of USDC loan yield in the ETH market. The table shows that higher ETH prices, volatility, and stablecoin supply generally correlate with lower USDC loan yields. At the same time, the stronger the MACD signal, the higher the yield.
Comparison with traditional risk-free rate of return
While it is interesting to uncover the reasons for the low-risk yields on stablecoin loans, it is also important to compare these yields to their counterparts in the TradFi market.
Because the stablecoin lending rate comes from the realized floating rate of mortgage loans on the Compound platform, we chose the General Collateral (GC) rate used in the traditional money market as a comparable risk-free rate, because it is also a floating rate with national bonds as loan collateral .
The figure below is a chart of the investment portfolio value of USDC loan yield, USDT loan yield and GC interest rate yield respectively. All investments start with an initial value of $100 on 2020-05-01 and end on 2022-05-01. As shown in the chart below, the yields on USDT and USDC mortgages are significantly higher than GC rates. On the other hand, risk-free investments earning GC rates have seen little growth over the same period.
The average interest rate in the table below also confirms that the GC interest rate averages around 0.08%, while the loan yields of USDC and USDT during this period are 3.71% and 4.51%, respectively, as shown in the figure below. (We also looked at the 2020-05 2-year yield, which is just 0.2%).
in conclusion
in conclusion
The paper presents a broad indicative analysis of the low-risk returns offered by stablecoin-collateralized loans via DeFi protocols. While these yields may fluctuate from day to day, their overall trend is better explained by BTC/ETH price, volatility, stablecoin supply, and MACD (momentum trading activity). We also compare these returns to the risk-free returns of the TradFi marketOriginal link
IOS
Android