Understanding QiDao's Risk Management Framework

Risk management has always been at the core of how QiDao operates, but as the markets have become increasingly more volatile, the use of a risk management framework for any deFi protocol, particularly a stablecoin lending platform such as QiDao, has become essential to daily operations. Thanks in part to our strict adherence to this risk management policies, QiDao has seen continued growth in the bear market, while maintaining user funds and the protocol safe.

The framework QiDao uses is based on Aave’s own risk management rubric. If you frequent our own collateral matrix at https://mai.watch, you may find the following rubric to be very familiar.

The reasoning behind the rubric criteria are fairly simple. As a collateral contract increases in age, it should also increase in marketcap and volume if the collateral contains solid tokenomics and general market adoption. The criteria are divided into three categories: smart contract risk, counterparty risk, and market risk. Let’s use the ETH matrix on Ethereum mainnet as an example.

In order to assess a collateral, we first begin by taking a look at the age of the contract, and the number of on-chain transactions. This gives us a maturity score for smart contract risk.

Next, we look at the number of token holders, not including minting or governance wallets, as well as the level of decentralization for the token. This gives us centralization and trust scores which are averaged together to gives us a counterparty score.

Finally, we look at a token’s marketcap, average daily volume over three months, and the normalized volatility score (based on closing price…more on this later) to give us a diversified average score for market risk.

We then average all three category scores together to get a final rubric risk score. Seems simple right? Not entirely.

First of all, all the metrics that go into creating a risk score are constantly fluctuating. When we first began performing collateral assessments, this was done manually on a spreadsheet, and it took hours to gather all the data for our dozens of collaterals. As a result, collateral scores were updated monthly. This served us well for a short period of time as we could watch market conditions and update more volatile collaterals accordingly, but it was not ideal. This changed with the introduction of the QiDao Risk Matrix website (https://mai.watch), and over the past month, we’ve implemented scripts that update much of this information for us on a daily basis with manual checks for accuracy. If market conditions change rapidly during the day, these scripts can be manually run to gather data and update scores accordingly. All market data is gathered from Coingecko and Defillama’s APIs.

Market volatility is a key metric to consider when assessing a collateral, and determining volatility manually is an arduous task. To find a collateral’s normalized volatility, we must first find the natural log between two closing prices, then calculate the standard deviation among a range of logs. We’ll skip the math here for simplicity, but we’re think of normalized volatility as the percentage movements in price for a token. What this looks like for users, the risk committee, and the team translates to our latest feature (releasing publicly next week) — volatility comparisons.

While most investors simply look for “price go up,” we want to assess collaterals based on “risk go up,” and so having an easy way to analyze volatility at a glance makes it easier for the community and the team to assess a collateral’s risk behavior across time (90 days for our rubric charts). Notice that all collateral volatility charts also display WETH and WBTC volatility data for comparison, as these are generally considered the “gold standard” of collaterals across deFI, making it simple to see how volatile an asset is compared to these two tokens.

Astute readers will notice that there has been no mention yet of other metrics we constantly discuss on Twitter and Discord such as liquidity and slippage. While the collateral rubric can help us assess a collateral’s safety and validity as a collateral, it does not give us the full picture. Since we are using these collaterals to serve as backing for our stablecoin, liquidity and slippage are key metrics to consider.

Because MAI’s peg partially depends on the underlying backing being able to be sold to repay loans (see our documentation on liquidations), it is critical that the collateral be liquid enough to be sold quickly and for low slippage. The general metric used for slippage here is “largest trade for stables with <10% slippage.” Liquidators are an important part of the system, and therefore it’s key that liquidations are profitable in order for them to happen.

When looking at debt ceiling raises, we also analyze these last two metrics in order to assess whether a particular collateral vault can support the additional injection of MAI it is being given.

Of course, now that we can track volatility essentially in real time, we should also be looking at this data for MAI and similar stablecoins.

Coming with the next public release of the risk matrix will be a stablecoin volatility chart within the dashboard page. We have chosen to showcase MAI’s volatility against stablecoins with similar backing mechanisms to MAI so you’ll also find volatility data for alUSD (Alchemix), LUSD (Liquity), MIM (Abracadabra), and sUSD (Synthetix).

While users often pop into the Discord to discuss how to get MAI back to $1, note that we have a soft-peg of $0.99–1.01 so while being as close to $1 is the goal, it is important to consider MAI’s volatility to be one of the lowest of any comparable stablecoin.

Less volatility = more stability. Thanks for reading!

Subscribe to Qi Dao Risk Management
Receive the latest updates directly to your inbox.
Mint this entry as an NFT to add it to your collection.
This entry has been permanently stored onchain and signed by its creator.