Skip to content

Latest commit

 

History

History
85 lines (68 loc) · 5.25 KB

risk_8.md

File metadata and controls

85 lines (68 loc) · 5.25 KB
Raw

Risk #8 - 8/30/2018

Introductions
  • Rich, Steven
Defaults in the system are a sign of a healthy economy. Is that correct?
  • Natural rate of unemployment: a certain number of people being unemployed is expected in a healthy economy
  • Balance between liquidation ratio and business competitiveness for providing leverage on collateral
  • The maker ecosystem is comprised of risk avoiders and risk seekers and without that risk (and the default) there would be no mechanism to provide stability
  • Individual defaults prevent overall default
Will there be caps on individual CDP sizes?
  • Initially no, but ideally there would be
  • The impact on liquidity is based on the size of outstanding collateral
  • We’ll rely on the overall defensive ceilings to prevent over exposing to one collateral type for now
Doesn’t this require monitoring every collateral type constantly? How often will risk teams be checking on models and collateral?
  • Yes it’ll be continuous monitoring
  • Dashboards would relay information to teams to keep them in the know
What happens when risk teams are reacting to events around specific collateral types?
  • If it is a material change in the status in the market then an emergency proposal could be brought forward
  • In general, the next actions would take place during the next executive vote
  • Governance should occur period to period, rather than knee-jerk reactions
If we have 3 distributed risk teams, what are they there to push forward? Where do their allegiances lie?
  • Their allegiances align with their individuals
  • MKR holders would have significant interest in creating many, many risk teams
  • Service providers could provide value and would be reimbursed in a grant program or other
How do we factor in the auction mechanism in the liquidation ratio?
  • The liquidation ratio is assuming that someone has already bought the collateral, so the auction
  • Liquidation doesn’t necessarily mean getting to USD. Could mean another more liquid collateral
Can [Emergency shutdown] apply to a single collateral type?
  • You could, but its intended to protect from existential or systemic risks
  • All CDPs being bitten would be the same as an isolated shutdown
Now we're talking more about building the buffer above 100% collateral to determine Liquidation Ratio
  • Given a specific period of time, what is the likelihood of a collateral falling 1% in 1 day?
  • How much could you possibly lose in a single day?
  • What is drift? —> the general direction is this token heading
  • In a bear market, the movement downward will be exaggerated
What does the parameter of the liquidation ratio attempt to control for
  • Liquidity risk
    • You’ve got XX collateral and you have 7 days to get rid of it. So you plan to sell 1/7 of it each day
    • Run up against the spread
  • The volatility risk is calculated off daily price, usually closing price which isn’t indicative of intra day volatility
Correlation risk
  • We’re not working with a traditional portfolio where we can use a risk adjusted return point of view. We’re seeking to maximize stability and integrity of the system.
  • Diversification of two different assets occurs by looking at the two assets volatility which have slightly different volatilities that provide a natural buffer.
  • Want to be cognizant of creating pockets of concentrated exposure around some collateral types

We’re building up a whole set of tools to manage the risk exposures of our portfolio.

Price feed risk
  • Largely unique to crypto markets
  • Most countries have 1 exchange, so there is very limited price discovery. Some have 2 or 3 exchanges. The US is unique and has dozens of exchanges.
  • How do we determine the correct representation of price?
    • A centralized exchange like Coinbase or Gemini?
    • Need some sort of consensus around a real source of truth
  • Median transform - (12- 15) oracles that bring in the appropriate prices and we take the median of those
    • Mean is vulnerable to statistical outliers
What is the difference between Price and Value?
  • When you look at an exchange you’ll see the ‘advertised price’.
  • Value is a more pure sense because of liquidity in the market. Given exiting large positions will include endogenous cost
Why are some countries limited to few exchanges?
  • First and foremost a function of regulation, a form of control
  • Barriers to entry for liquidity and licensing are generally high
  • If an exchange is run by the state then there would be extended risk
How do we weight or apply trust values to the data coming from exchanges that we’ll rely on for price feed data?
  • If we’re looking to exclude wash trading, we can apply a simple proxy: daily volume vs market cap
  • Apply this proxy while assessing trading volume during asset qualification
  • Need to make a distinction of the issue from the quality of the assessment
How much are the MKR token holders on the hook for is something goes wrong?
  • Economic capital - expected short fall, we’ll explore more next time
What are some of the existential threats to the maker system?
  • Stupidity, manipulate the parameters of the system to crazy
  • Landscape, if Ethereum stops, then we’d need to adapt to that