Skip to content

Episode 38: June 6, 2019

Agenda

The theme for this call will be 'Collateral Risk'

  • 00:00: Intro from Rich Brown
  • 16:33: Analysis from Vishesh
  • 48:11: Intro to Collateral Risk with Cyrus

Introduction

Rich 00:00

  • Don't be hesitant to give us feedback, questions, comments, and concerns about these calls and conversations!
  • This is not a governance committee, we don't decide things on this call. This where we debate things, talk about trends, clear up misconceptions, present analytics, and so on.
  • We are shifting the focus away from the peg and more towards discussions around Collateral types and Risk.
Governance 04:00
  • We have a voting system that we are trying to make accessible to all potential MKR voters.
  • How do we make voters feel like they can make a difference?
  • How do we communicate most effectively about Governance events and current topics of discussion?
  • How do we keep people engaged in the process?
  • How do we communicate to voters how they can best contribute to the Governance process?
  • How do we get more MKR holders to use their tokens in the voting system?
    • The system is not as secure as it should be, we need to get more MKR staked in the current Governing proposal. Otherwise, unexpected changes can be passed with lower amounts of consensus.
  • What we're already doing to help fix governance:
  • This call is just the first of many places where information gets disbursed. We're working on more calls globally, in order to make this format available to others around the world.
  • We are providing summaries of this call for those that can't listen to the entire call or join during.
  • The voting dashboard version 1.5 is coming out soon. This will provide us more flexibility in the Governance polling.
  • Working on a hub for resources to inform voters.
  • Coinbase enabling MKR voting through their custody solution.
  • Things we need:
  • Push notifications around Governance actions.

Cyrus 13:42

Governance
  • Us having a high cadence with polling helps mitigate any mistakes. Since, if some crazy vote goes through, we can quickly undo it.
  • Assuming we slow down the cadence, it would be interesting to see if we can create some emergency response drills or procedures.

Risk

Vishesh: Analytics: 16:33

Vishesh's Graphs DAI 24hr VWAP Graph

Dai Price
  • Coinbase has been added to Dai.deciepher.io
  • The Dai VWAP has been distributing on a wider spread over the past few days. This is something to watch.
  • The Dai price has been doing well, there has been some variation dipping below $1. But on the broader 1-month time-scale, it's been healthy and on target.
  • The situation has changed slightly. When Dai dips below a dollar now people are more likely to take this as a clear arbitrage opportunity.
  • The Stability Fee was just changed to 16.5% but it's too early to see the effects.
  • Dai/ETH price trend has reversed as we've been mentioning in previous calls. It's no longer the case that if ETH price is rising than Dai price is falling.
  • It would be interesting to examine why ETH has begun to track Dai price really well.
Dai Supply
  • Supply ticked up a bit over the past few days. Both significant draws and wipes occurred, but overall draws prevailed.
  • It seems like refinancing has slowed down.
  • The rate changes may have taken the pressure off of supply dampening.
  • The circulation of debt has held fairly steady. Open debt is younger than closed debt, meaning we are eating into the backlog of old debt at a steady rate.
Collateralization Ratio
  • Collateralization Ratio has come down, from 540% to ~460%.
  • This lines up with what's going on with Dai price and supply. The Dai supply coming up with some significant mints, along with a looser Dai peg, tells me there has been some uptick in leverage-taking.
  • Decreased Col. Ratio and increased Dai supply may indicate that there was some uptick in leverage.
Questions to Vishesh
  • 26:04: How long does it take for us to see cause and effect in response to the Stability Fee changes?
  • My short answer is that there is no defined time-period. I do think there is a certain minimum level of resolution that comes into play after 3-5 days. However, there is a huge element of guesswork involved in approximating this range.
  • The bigger thing is how do we judge the time-scale when there are multiple other changes occurring: for example a SF decrease paired with a ETH price drop.
  • 33:52: Can we assume that, if the peg seems stable, should we be lowering the stability fee as a best practice?
  • It is hard to know the motives of MKR voters. However, if the system can function properly while having the SF becomes less expensive, that is probably a good thing.
  • It would be interesting to see the overlap between MKR holders and CDP users.
  • 36:33: Isn't the drop in the Collateralization Ratio a result of a falling ETH price?
  • Mathematically, yes. But what we've seen historically is that within 2-3 days after such a downward price movement people tend to move the collateral ratio back up. However, this time it has not happened and has continued to go down despite ETH price continuing to drop.
  • 38:22: Discussion on DSR effect on Secondary Lending Markets

Matthew Rabinowitz 46:38

  • Alex and Vishesh are continuing R&D on the PID idea.

Cyrus 48:11

Collateral Risk Intro & Discussion
  • Many of these discussions take place on Rocketchat, Reddit, and other forums. It will be difficult to give a full overview of Collateral Risk on this call, but we can try.
  • Today I will be covering a few different things:
  • Overview of various aspects of Risk.
  • Community involvement, and how you can contribute.
  • Challenges for some of the risk models.
  • High-level topic: What are we solving for? What are we concerned about?
  • Understanding how CDPs can be modeled is probably the most crucial aspect of the entirety of collateral risk.
    • CDPs are loans that people issue to themselves based on the collateral they use.
    • Not based on credit risk at all, but on collateral risk.
  • The first step is defining exposures; who is exposed to what types of risk?
    • The exposure risk of the CDP comes down to the probability of default. Meaning, if it does default how bad the loss would be.
    • How much of these losses would result in bad debt (aka debt that requires recapitalization through the issuance of MKR).
    • Essentially the two topics are: What is the likelihood of CDP default & given that number, what's the upper bound on the exposure we can facilitate (which is defined by the debt ceiling).
    • The stability fee that we attach to a type of CDP will compensate for this risk.
    • In order to assess this risk, we can go a number of different routes. We can start with academic frameworks, and then work our way down to pragmatic solutions.
  • 01:03:41: Qualitative Assessments:
  • The Qualitative assessment helps you identify risks outside of ones you can observe from the market history of the asset.
  • This can be a good jump-off point for the community to contribute.
  • 1st consideration: How does that collateral appear on our doorstep. We need to come to some agreed-upon method for an onboarding approach. Determining a priority would include factors such as:
    • Safety of collateral
    • Amount of Dai potentially generated
    • how uncorrelated it is from the rest of our portfolio
  • The next steps would be to classify the collateral, perform due diligence, and finally a risk analysis.
    • One of the most important determinations is whether potential collateral is really an asset or not. Meaning is it a bearer asset? does it hold counterparty risk?
    • From this point, the next step is Due diligence which requires careful examination of the collateral asset. Looking at the team, the technology, the business logic behind it.
    • Risk analysis tries to examine factors that can lead to bad market conditions. We want to uncover anything might cause an immediate loss in collateral value.
Questions for Cyrus (Also counts as post-call questions)
  • 53:00: How do we quantify or track the basis of these exposures?
  • Theoretically, that is what we would like to do. In practice, we will experiment with a number of approached. We will likely find the need for programmatic solutions to help.
  • 58:08: How are you reasoning about the Risk premium? What is its purpose and what factors into it?
  • Compensates for bad CDP liquidations(Liquidations that further require recapitalization through the issuance of MKR).
  • 01:00:56: How does the system respond to bad debt events? Is the Collateralization Ratio raised if an asset experiences bad debt past a certain threshold or at all?
  • The problem is the risks that don't necessarily manifest in the history of the market for a certain asset. Some bad debt or severe market events are hard to account for.
  • 01:05:50: How do we reason about Collateralization Ratios versus Risk Premiums?
  • The Collateralization Ratio is a metric to measure against expected maximum slippage during the auctions.
  • The Risk Premium has more to do with the Risk of many CDPs of one type becoming subject to liquidation at once.