Weekly (almost) Narrative on MakerDAO - 28 July 2019

General:

During the course of the last week, the community held a polling vote where the winning proposal was to decrease the rate by 200bps to 18.5%. At the time of this narrative, the executive vote looks somewhat unlikely to pass as an edge case with the voting contract was discovered / stumbled upon. (Detailed summary may be found at An Explanation of Continuous Voting and the Peculiarities of the 7/26 Executive Stability Fee Vote ). That said, should the conditions that caused that edge case resolve themselves, the proposal / spell to reduce the overall SF would pass thus reducing the SF to 18.5% per annum.

The overall price* of DAI has solidified around its soft peg target of 1.0000 and for the most part stuck to the peg with times being above 1.00000 . During the last week, the total outstanding DAI has largely contracted and only recently started to increase to 76.3mm DAI. Several large CDPs have been closing out causing a large contraction in the DAI outstanding. This on the surface could be construed as a negative signal. However, when viewed from a macro lens and watching the Compound utilization increase, what appears to be occurring is a large scale refinance, which is healthy to an economic & monetary system. Further, with the recent release of the instadapp refinance tool, this narrative is further confirmed as large quantities of DAI were migrated to Compound. It is expected that tools like instadapp will continue to refine their refinance capabilities to expand to additional secondary lenders

Over the course of the last months, the DAI ecosystem has been addressing an increasing supply and trying to keep supply and demand in harmony (thus having a DAI peg). As we have perfect visibility into the outstanding supply at any point in time, we struggle and continue to struggle to identify the actual demand. Tools like instadapp that allow for efficient refinance help reduce excess supply and pull the supply down to the true demand. Post MCD, the DSR will have a similar impact and should be equally or more sharp.

As DAI can be created and destroyed only with Maker, this refinance process will continue until the utilizations on all secondary market lenders hit 100% and their lending rates are at parity or more than with a Maker CDP.

During the course of the last week, we saw the Compound utilization almost hit 100% and the dydx lending rate is now higher than DAI minted directly with Maker.

This is critically important as it now points to new DAI being minted is coming from core demand. Earlier in the year during part of a bull market (for the core collateral, ETH) we saw the supply increase and there were calls to increase the debt ceiling. More importantly that surge in DAI caused the price to be degraded below $1.00*. Retrospectively, the market had excess DAI and was subsequently moved over to secondary markets. Further, the DAI supply could be constrained by increasing the SF to the point of supply destruction thus restoring the peg.

Today, the market is far more consolidated and more importantly it has sustained an elevated SF and has saturated the secondary markets to the point of almost 100% utilization where their lending rates are more or less on par with Maker itself AND now the price is still holding at the peg or even slightly higher.

This means that excess supply is not driving the ship, rather core underlying demand is. When this occurs, to not have the peg break upward, the community will be pushed to decrease the SF to cause the issuance of new DAI to fill that new demand. This is a wonderful sign and at the same time a yellow flag.

The Maker / DAI ecosystem has never been this size and had demand in the driver seat. Further, the market is going to start the process of forcing rate decreases to address the peg. There is no demand elasticity data on how much DAI will need to be minted to satisfy this demand. Further, the pace of that new minting is also an unknown. With a debt ceiling of $100mm, the breathing room needed while also voting on other community items surrounding multi-collateral DAI becomes quite tight.

In the face of this situation, it is recommended to form rough consensus on what would cause the debt ceiling increase and to what degree such an increase should be warranted.

During the past week, the Maker Foundation released updates related to the MCD launch. While the exact launch date is unknown and the MKR token holders must vote to officially launch the MCD smart contract, there are many governance actions the community should be taking now in advance of such a launch.

MCD Collateral Package:

Most notably will be the new Risk Premium (associated to ETH) needs to be determined and a plan to deploy that new Collateral Package on MCD. While the interim risk team will recommend the specific risk premium, the market is already signaling a rough estimate by taking the difference between the borrowing lending rates on Compound or dydx for ETH. This also implies the DSR at present for DAI (with ETH as the sole collateral) will need to more or less equate to the lending rate at Compound / dydx). Thus if we estimate the difference between the borrowing and lending rates would be the needed Risk Premium (or even less as the collateral package at Compound is more “risky” than Maker with a lower collateral requirements), we can roughly assume the risk premium allocated to ETH will be in the 6% area.

Given the drastic difference between the current SF of ~20% and ~6%, the community needs to have rough consensus on a deployment plan to launch a collateral package that is more or less the same that exists in SCD for MCD. Such a proposal has been suggested on MakerDAO forums but has yet to receive rough consensus or a vote from the community. (Navigating the waters between now (SCD) and MCD with the DSR)

Voting:

During the last week, the edge case (referenced above) appeared. In it a possible system vulnerability was revealed. This vulnerability is not an attack vector per se or one where funds are at risk, rather one where MKR committed to a specific legacy proposal / spell could inhibit future proposals / spells from being implemented / cast. For a system that is designed for the “long-term” that easily could run for decades (or more), the human aspect of mortality must be considered. At present, should a MKR holder vote for a legacy proposal / spell and not move his / her MKR, it presently somewhat acts like a blocking tool for future proposals / spells. While well intended, it is strongly recommended to implement a proposal / spell expiration if not implemented / cast within X days. That X should either be a fix number that has rough consensus or one that the community can change as needed. Today, this edge case is manageable; however, as more MKR is burned as a function of operations / time, the more this risk gets magnified with time unless it has expiration implemented (similar corollary to auto deleting email for an enterprise after X days, not an issue when first implemented, however if left unaddressed the issue magnifies with time). As voting is essential to how MKR functions, this issue is particularly acute. There have been several reports that this issue can be solved with minimal changes. As such, it is strongly recommended to implement those changes before apathy sets in. Vote and vote often should be a common theme. The community simply cannot risk legacy MKR being allocated to a possibly stale proposal / spell and having the owner of that MKR being unavailable to move said MKR for reasons which might include death and that inaction possibly holding the voting of the system hostage.

Terminology:

At present, the only tool in the toolbox to increase or decrease supply is the stability fee. While the nomenclature is commonly referenced, it is important to take a moment and point out how the stability fee in SCD and MCD are drastically different.

SCD:

Stability Fee(SCDx) = RP(SCDx)

MCD:

Stability Fee(MCDx) = DSR(uniform) + Oracle Fees(MCDx) + Risk Team(MCDx) + RP(MCDx)

As such, there is a fundamental shift in the thinking about how the community strives to meet its core objectives of maintaining the soft peg of 1.00000 pre- and post- MCD.

In SCD, the objective was to control the supply by modifying the SF(SCDx) to find the equilibrium of where supply would meet demand thus by derivative the price would meets its soft target of 1.00000 That is to say the SF(SCDx) was being used to create incentives for supply increases or decreases.

In MCD, the objectives shift and are both magnified and segregated. The community must continue to meet its core objectives of maintaining the soft peg of 1.00000 via supply and demand but now must also manage risk in a way not done in SCD.

In SCD, the SF(SCDx) was only used to control supply. In MCD, the RP(MCDx) is used to control the risk of a given collateral package, in theory ignoring completely the subsequent supply created as long as it is risk-adjusted supply.

For reference, a collateral package is being defined as any given collateral that has different parameters. Those parameters include debt ceiling, liquidation ration, and collateralization ratio. A given collateral may have one or more of these packages. Each of these packages when viewed compared to the system and the collateral itself must be initially priced via RP(MCDx) to the point where DAI that is minted from that collateral package has been appropriately de-risked to the point of being “almost” riskless. While nothing is every truly completely riskless, the objective remains the same.

Risk diversification:

However to have a truly global tool that underlying portfolio of collateral needs to diverse and as uncorrelated as possible. That is to say that credit concentration becomes a concern. Thus certain market forces are needed to both encourage and discourage certain behaviors that will be applied to that RP(MCDx). For example, should the RP(MCD jpm debt) be determined to be 100bps, that must be viewed from the perspective of status compared to the portfolio as a whole. Thus if the only collateral that is used is JPM debt and it is viewed as stable, the initial RP might be given a discount to encourage the onboarding. Further, other collateral types may be given a discount to ensure a well-rounded portfolio. Following on, if a given collateral package is utilized too much, in addition to the debt ceiling a RP penalty may be utilized to increase the borrowing costs to discourage DAI creation from collateral that might be too concentrated when viewing the portfolio as a whole.

In the last week, Fluidity presented / announced their intent to harness the credit worthiness of a US Treasury as collateral (https://medium.com/fluidity/introducing-the-tokenized-asset-portfolio-7710e4239ab6). Such an initiative is essential to the growth of Maker’s collateral pool and represents an essential growth plan for MCD, not only for DAI issuance but risk diversification and bringing uncorrelated assets onboard.**

Maker Education:

The process of learning about this system is rarely linear. As personally witnessed / experienced and published via this narrative, several positions have had opposing iterations as more revealing information was applied to the core understanding over time. The “I have a stupid question” fear is absolute poison in a community driven system like MakerDAO and must be constantly addressed / battled. There are no stupid questions, just stupid people. Asking and answering question removes that ignorance for the broader community. The removal of ignorance by direct vigilant application converts stupid people to well-informed people. Further, as we are all learning, it is recommended at least once a month to have part of the Tuesday Maker Community call allocated to Ask Me Anything about Maker and help folks answer any questions they may have. To that end, a broader article / book is being authored in the background to help (MCTT). More on MCTT when it is further along.

General theme:

The community needs to have rough consensus not only where we are going, but also on the short-term big picture steps on how to get there.

    • price being determined by USD fiat offramp via USDC - DAI (at pro.coinbase.com)

** - full disclosure my team and I will be advocating a similar cross-collateral platform focused on commercial real-estate and other credit (investment grade or other) backed projects in the near future

NOTE: Not a part of the Maker foundation, just my $0.02 and not intended as advice in any capacity.

2 Likes