Dai - Solving the Stablecoin Drama the Decentralized Way
Crypto Monday, March 2019
Stable coins
“You need stablecoins to realize the full potential of blockchain technology.”
Why?
- Augur bet in 6 months
- ICO participation (both sides)
- Remittances
- etc.
→ Majority of use cases require a stable coin.
Basic requirements
- Pegged to a reference value ($, €, IMF SDR, …)
- Backed by something to assert value
Risks
- Risk #1: Irrational market 🙃
- Risk #2: Counter Party / Collateral 😕
One Solution
Fiat-backed stablecoins are IOU tokens and are usually backed by a bank balance.
Examples
- Tether (USDT, 2.000 Mio US$)
- USD-C (Circle/Coinbase, 239 Mio US$)
- TrueUSD (TrustToken, 198 Mio US$)
- Gemini Dollar (65 Mio US$)
- PAXOS, JPM Coin (JP Morgan), …
The Risks
- Risk #1: Irrational market 🙃
- COIN < USD:
- Arbitrage opportunity because of IOU redemption (Redeemability)
- COIN > USD:
- Arbitrage opportunity as long as you can buy/mint
- COIN < USD:
- Risk #2: Counter Party / Collateral 😕
- Blind trust, or
- Periodic audits to prove solvency of the IOU company
March 14, 2019
Audits
- Third party audits a remedy?
- Tether: 2017 no audits, 2018 claims to have been audited by sketchy firms
- TrueUSD: Monthly audit by Cohen & Company, a top 50 U.S. public accounting firm
- Trust is deligated. Meh.
Is there a better way?
Dai
- ERC20 Token
- pegged to USD
- launched Dec 2017
Idea
Provide stability by
- Collateralized Debt Positions (CDPs),
- autonomous feedback mechanisms, and
- appropriately incentivized external actors.
Sounds fancy, how does it work?
CDP
- Smart contract (think verifiable vending machine)
- Automated way to create and repay loans
- A smart contract can only operate on data that is on the blockchain
Most important slide
| Paid in | Collateral | |
|---|---|---|
| Bank loan | Euro | Your house |
| CDP | Dai | Ether |
Can we deal with the risks in a decentralized manner?
Dealing with the risks
- Risk #1: Irrational market 🙃
- NOT redeemable!
- Arbitrage (Keeper bots)
- Dai < USD:
- System incentives repayment of CDPs
- Dai > USD:
- System incentives creation of CDPs
→ Basically supply is controlled by incentives
Dealing with the risks
Risk #2: Counter Party / Collateral 😕
- On-chain “staking” of collateral
- Always over-collateralized (“Secure debt”), because of volatile nature of the asset
How much Dai can I draw?
- The amount of Dai depends on the price of ETH
- The ratio is called “collateralization ratio”
- Must be higher than the “liquidation ratio” (150%)
If Ether goes up
- Not a problem, collateralization ratio goes up
- Dai gets stronger as there is more collateral
- Payback amount in Dai does not change
Use Dai to buy more of the collateral asset.
→ Decentralized margin long
If Ether goes down
- May become a problem: A CDP can become “under-collateralized”.
- Forced liquidation when hitting liquidation ratio (Margin call)
Liquidation
- Auctions part the your collateral to the highest bidder
- Adds a penalty (13% of borrowed assets) to incentivize high collateralization ratios
Returning your Dai
- If you want to access your collateral, you have to close repay the borrowed Dai.
- Pay stability fees in relation to the holding period
All is good!
Is it?
Governance
- Who sets the collateralization ratio?
- Who sets the stability fee?
- Who white-lists the oracle price feeds?
- Who decides which tokens can be used as collateral?
→ MKR holders
MKR Token
- Governance (https://vote.makerdao.com/)
- Insurance against black swan events
- Receive a compensation
- Basic idea: “Share risk and profit”
Global settlement
- Last resort shutdown
- Everyone can claim the net value of the collateral they are entitled to
- Can be triggered by whitelisted MKR holders
Stability Fees
- Stability fees are collected in MKR and send to a burner contract
- By shrinking supply, MKR holders are rewarded for taking on the involved risks
Community
- Friendly community
- Weekly Community Calls / Governance Calls
“We are dedicated to providing material for new people to understand the system in depth. This will be important for successful governance in the project’s future.”
Symbiotic relationship: Dapps need Dai, MakerDAO need Dapps Medium post
Ok, but does it work?
Scale
- USD 4 Million in BTC Lightning Network
- USD 300 Million in CDPs (~2 % of all ETH)
Dai Supply and stability fee
Where we at?
- Tether (USDT, 2.000 Mio US$)
- USD-C (Circle/Coinbase, 239 Mio US$)
- TrueUSD (TrustToken, 198 Mio US$)
- Dai (95 Mio US$)
- Gemini Dollar (65 Mio US$)
Discussion
- Use cases and related projects
- Risks and criticism
- Future developments
- Wrap up / The big picture
xDai Chain
- Ethereum side chain built by POA Network
- Proof of Authority, 5 second block time
- XDai native token (representation of Dai)
- Locking ERC20 tokens on Mainnet will mint those on side chain
- xdai.io Burner Wallet
- Powered food trucks on ETHDenver 2019 (~35k USD turnover)
Criticism 1/2
- Deflation risk of the collateral (“CDPs swallowing ETH”)
- Failure of governance
- Regulatory uncertainty
- Who fights the legal wars? InstaDApp or MakerDAO?
- Taxes?
Criticism 2/2
- Inability to scale to a global stable currency?
- Might not be able to meet demand, while keeping over-collateralization
- Price Wars with non-decentralized projects
- Tether could clone the system and offer 0 % loans
- Leverage up to 1.67 % / BitMEX up to 100%
Alternatives
- Algorithmic Supply Stablecoins
- Metal-backed Stablecoins (DigixDAO)
- Other
Future developments
Multi-Collateral Dai
More stability and scalability through diversification
- ERC20
- BTC
- Tokenized Bonds/Stocks/Art
- …
→ Any collateral could be valid. As long as sensible parameters are chosen.
Next level
- Clone contract and use Dai as collateral
- Ability to create synthetic assets:
- Synthetic gold
- Synthetic stock index
- Leveraged margin short
- “Not insured” by MKR holders
The Big Picture
Triffin Dilemma
- National currency as world reserve currency
- Controlling supply and demand of own economy
- vs. providing enough supply for the world
- This dilemma killed the gold standard of USD
If you squint, MakerDAO looks alot like a central bank.
Yea, but…
- A sound, transparent, non-fractional-reserve bank
- Supranational: Not bound to any specific economy
- Restricted by design: Code-as-law
To overcome the Triffin dilemma, we need a stable currency which is governed globally and independent of any specific state or economic area.
Sound familiar?
If interested, watch this talk: MakerDAO: A New Hope for Financial Reform
Links
- https://makerdao.com/en/whitepaper/
- Purple Paper - Reference implementation in Haskell
- MKR Tools
- MakerScan
- https://www.daiembassy.online/
- https://github.com/makerdao/awesome-makerdao
Bonus slides
Oracles
- “The smart contract needs to know the price of ETH”
- “A smart contract can only operate on data that is on the blockchain”
- Oracles to the rescue: The median value is chosen from multiple Oracle price feeds.
Oracle contracts
- Invididually owned Price feed contracts will be called periodically and store the value.
- Then they
pokethe Medianizer contract - On each poke, the medianizer calculates and stores the median price of all feeds.
- Governance can whitelist certain price feed contracts.
The cool thing is
…other contracts can call read on the Medianizer.
Other Dapps like Compound, Gnosis and Augur are using it.
Room for improvement
- n price feeds means n times the gas every update.
- Possible mitigation (Oracles V2) is to instead of calling a store and poke function on-chain, sign+send a message to a relayer who will then post it on-chain
ETH, WETH and PETH
ETH → WETH → PETH → Dai
- WETH: Wrapped ETH (ERC20 representation of Ether)
- PETH: Pooled ETH
Short technical dive
open– initialize CDPlock– lock up collateraldraw– issue Daifree– reclaim collateralwipe– send back Dai (reduce issuance)shut– close CDPgive– transfer CDP to another ownerbite– trigger liquidation / apply penaltygrab– receive CDP’s collateral while in liquidation
