- Rich, Steven
When should the stability fee be increased or decreased? What is our point of reference (outer bounds)? [1:48]
- The Dai exchange rate will be a heavy indicator as to the policy being adopted around stability fee
- Supply and demand fracture also contributes
- Look at a combination of exchange rates and trigger a force of action
When it comes to data we should look for exchanges that have aggressive wash trading policies. How much backtesting should we consider before establishing a range of exchange rates? [5:54]
- The reliability of the data will drive the efficacy of any empirical analysis based on that data
- While we are bootstrapping we have to work with imperfect data
What models are we relying on right now to draw these empirical conclusions? [8:33]
- Largely abstract models during bootstrap, looking at data points and understand what insights they can provide
One of the factors is considering the natural arbitrage opportunities of market making teams and how the pressures affect them as the peg is drifted [13:41]
- Even though market makers will tend not to reveal all information to remain competitive they could contribute to their desired stability fee rate through a rate like LIBOR in the traditional world.
Can we trust market maker's analysis of what the stability fee should be? [15:10]
- On it’s face, no but they should have self serving interest so we can view it in that light
So we want to have a signal that allows us to rigorously say that if x happens we want to come together and consider y action? [18:44]
- The implementation is still being bootstrapped but that the’s general idea
- It could be a matter of the community taking initiative and putting together proposals for adjusting the stability fee by a quarter of a point or half a point
It appears the change of the stability from 2.5% to 0.5% has had the intended affect given data on the exchange rate since the change. Is that fair to say? [20:55]
- It’s tough to call causation rather than correlation
- It could be an anticipatory effect, but we should look to collect more data for empirical analysis of future decisions
We should also consider what we want the models to achieve for us [25:36]
- Sensitive to mismatches in supply and demand
- Be cognizant of the inventiveness of MakerDAO in the context of the broader ecosystem
- The stability fee should be utilized to support the system itself and needs to be allocated appropriately for grants, Dai Savings Rate, etc to spur growth in the right ways
What happens if the fee is set too high or too low? How do we know that has happened? [34:35]
- Since we don’t have the breadth and depth of feedback we don’t really know when it has happened
- What is an increment we’ve seen before, how did the exchange rate react to those changes?
- Where is the ceiling? 3.5 - 4% when no longer being competitive
How do we know that Dai isn’t trading where we want? [37:10]
- We need to put proxy metrics in place
- This meeting can help establish those practices
- We can solicit from the community what types of things they’re looking for in good data sources
- Establishing how the stability fee affects —> creating risk teams —> and compensating those teams
Is there a case for a feedback-loop controller framework? Will there be an algorithm to determine the appropriate stability fee? [43:24]
- We’ll need a robust algorithm from the information analyzed
- The algorithm would remain a recommendation while governance is the last say. Eventually that could change
It’s becoming apparent that the stability fee is a classic variable rate. Is there friction in describing it this way and setting the expectation that it can and very well might change? [46:24]
- At critical mass in the traditional world, a few basis points can have a large effect
- As we’re bootstrapping the system the movements are less detectable on the intra-percent level
- Fixed rate interest has benefits for very large CDP holders or risk seekers
We’ve adjusted the fee in 2% increments so far. Should we consider smaller movements at 0.5 or 1%? [49:32]
- We should consider that and the cadence in which votes are considered or at least what the “normal” looks like
- A general suggestion would be around 0.5% at a time
- We could continue having weekly discussions or place for contributions. We have open chats that any one is welcome to share ideas in those channels
Closing thoughts: We should consider raising the stability fee by 0.5% [55:38]
- We’ve moved at a 2% change and seen the effect. We could make a smaller change and consider the further effect
- Please raise feedback through reddit, rocket chat or on the next risk call