Beyond the hype on Maker DAO — eliasimos.xyz - Cryptoassets, Data and Markets

Part 1

For the next few editions I'm going to dig into DeFi's most prized asset - Maker DAO. Being cognizant of time limitations here, I'll break this up into multiple pieces - which will hopefully make it a little bit more digestible for all. Maker has conjured up a lot of excitement among the community and for good reason, as at the moment around 2% of all the ETH in circulation is locked up in the Maker smart contract.

What exactly is Maker though? In short, Maker is a set of smart contracts responsible for the issuance of the DAI stablecoin (pegged to the dollar). DAI is backed by Ether that is locked into the Maker smart contract as collateral - see below for a glimpse of the UI.

As an ETH holder, you can lock up your ETH as collateral and issue an equivalent amount of DAI. This function is known as a “collateralized debt position” or CDP. Due to the volatility of ETH, the amount that needs to be provided as collateral is a multiple of a 1:1 relationship, to prevent positions being liquidated on wild price swings.

In the event of a reduction in value of the collateral asset (ETH), Maker liquidates CDPs and auctions off the ETH inside for DAI until there is enough DAI to pay back what was extracted from the CDP. So far, so good...

What happens though, if the price of ETH crashes well below the one-to-one collateralization ratio in a very short time-frame. This is where MKR - the native token of the ecosystem - comes in. MKR is an ERC-20 that has governance rights over the Maker smart contracts. As mentioned in a previous version of the newsletter, the governance rights granted to MKR holders, preside over state; put simply, holders of MKR have influence over the key risk parameters of the system in a way that materially changes the composition of the system (state). These include:

The Debt Ceiling - maximum amount of debt that can be created by a single type of CDP
The Liquidation Ratio - ration between collateral and debt
The Stability Fee - an annual percentage yield and needs to be paid by the CDP user
The Penalty Ratio - the max amount of Dai raised from a Liquidation Auction

Incentivization to participate in the governance of the platform as a MKR holder, materializes in stability fees. The more collateral in the system, and the more DAI is issued, the higher the value of those fees. As it becomes apparent, governance rights are valuable in the Maker ecosystem. We'll dig deeper into exactly how much tomorrow.

Part 2

Today, we continue with the dive on Maker from where we left off - governance on the platform vis a vis the MKR token. The way governance materializes, is via proposals instigated by different members of the community (which is largely organizing in the Maker DAO subreddit), that are voted for by staking the MKR token. Voting power is equivalent to the amount of MKR at stake.

The voting process for the governance is done through continuous approval voting. In simple terms, any MKR holder can vote for any number of proposals with the MKR that they they hold. They can also submit a new proposal or cast or withdraw their vote any time they want. The proposal with the maximum votes from MKR holders becomes the top proposal and can be activated to implement any changes in MakerDAO.

The incentives interplay for MKR holders and governance participants is two-fold (at least on a high-level); participating to earn the stability fee, but also participating to ensure that the platform is well functioning so that the long term value of the investment (MKR) is more likely to increase.

This is an interesting relationship, as often the governors will be called to sacrifice short term gains (less stability fee) for a longer term payoff (MKR price appreciation). Under the MakerDAO system, stability fees are denominated in DAI but must be paid in MKR. Once paid, the MKR is burned, reducing the supply and hypothetically increasing the price of MKR. Reducing stability fees means that less MKR is burnt, potentially decreasing the price of MKR. We actually such a turn of events play out many a times in 2018, with the latest event recorded in December. Governance requires constant involvement, in order to ensure that it is always a representative number of votes changing the state of the system, and not the other way around.

As you can imagine, when such critical decisions are at - well - stake, consensus via a widely disparate demos can be hard to achieve, and maybe in fact not optimal in the long run - this is an opinion espoused by Rune Christensen, founder of the Maker DAO platform, in a recent edition of the Unchained podcast. Thus far, the constitution of MKR holders seems to be highly concentrated, mitigating that risk on the one hand, but on the other leaving a large amount of decision making power to the Polychains and a16z's of the world.

Santiment recently ran an analysis of the on-chain Maker DAO network and uncovered that there is indeed a high level of concentration in address interactions. While most ERC-20 contract activity is distributed as in the Santiment Network's case, in Maker the network of interacting addresses is a lot more tightly knit.

Another property that describes the governors, and arguably favours relative centralization, is their function as the buyer of last resort. Should the collateral in the system not be enough to cover the amount of Dai in existence, MKR is created and sold onto the open market in order to raise the additional collateral, providing a strong incentive for MKR holders to responsibly regulate the parameters at which CDPs can create Dai, as it will ultimately be their money on the line should the system fail. As such, deep governor pockets act more as a feature, rather than a bug.

In order to prevent system failures (e.g. a 51% attack or a sharp drop in the price of the collateral asset), Maker attempts to mitigate the risk in two ways; (i) a delay mechanism and (ii) a global settlement mechanism.

The delay mechanism kicks in after a vote is completed, in which time window stakeholders can audit the new code for 24 hours. If the code is reviewed and considered detrimental to the system, then the system will be shut down before it can be deployed.

Now in the event of a black swan event that critically threatens the system, Maker DAO initiates global settlement; a process through which the entire system freezes and all holders of DAI and CDPs are returned the underlying collateral. Global settlement can be triggered by a select group of trusted individuals who hold the global settlement keys. The collateral owners are at that point exposed to whatever market volatility ensues, but are nonetheless guaranteed return of their collateral, after which point the system reboots.

If an attacker bought 51% of the active MKR stake, and somehow tried to corrupt the system, the settlers would initiate the global settlement process. After that is complete, the reformed Maker DAO would be spawned and the attackers 51% stake would have been removed, leaving only the honest MKR holders left. Given both the carrot and stick type incentives outlined here, it becomes evident that corrupting the system is very expensive and therefore not an optimal play.

In pt.3, we'll take a look at how DAI keeps its peg, how that is likely to be affected with the introduction of multi-collateral DAI and what the general implications are.

Part 3

As outlined in parts 1 and 2, the way MakerDAO system works is that users pool Ether together (referred to as PETH, or pooled ETH), and after this pooling, the issued DAI tokens are collateralised against this asset-locked reserve. The way DAI maintains its peg against the USD then, is by adjusting the key parameters of the system to incentivize demand and supply forces to bring the peg back to balance. Consider the following example, as outlined by Alex Larssen; as the price of ETH crashed along with the rest of the market back in November 2018, Maker DAO triggered a series of CDP liquidations - a process that involves withdrawing DAI from the market and returning PETH to the CDP holders.

In the period examined, there were 3 main liquidation events, after which the DAI ecosystem found itself in an excess demand condition. As such, the governors brought the stability fee down (cost of creating a CDP) in order to make it more attractive for participants to provide collateral. Subsequently DAI returned to its target range. As you can imagine, in a period of excess supply (such as the current), the governors will strive to increase the stability fee, to part of the market to withdraw collateral and contract the supply of DAI (in fact there is a currently a vote in place to increase the stability fee to 1.5%).

The problem here, as far as stablecoins are concerned, is that there are no evident demand-side ways to keep the peg in place. As Su Zhu and Hasu explored in their recent piece, "Maker Dai: Stable, but not scalable", Maker DAO does not allow a margin for professional arbitrageurs to get involved in exploiting price discrepancies to bring the DAI peg to balance, primarily due to the fact that DAI is not interchangeable with its collateral on a 1:1 ratio. The following passage sheds some light into this;

When market demand pushes the price of DAI to $1.02, you can again take $1.00 USD, buy $1.00 ETH (or any other asset that can be used as collateral) and lock it in a CDP. The problem, however, is that for each $1.00 ETH locked up, Maker will give you less than $1 of Dai. That is due to the requirement for over-collateralization. The current collateralization ratio is 150%, so $1.00 ETH in a CDP can generate up to 0.66 Dai (this ratio could change, but it’s never going to be close to 100%).
While you wait for the price of Dai to drop, you are stuck in an unfortunate position:
1) You don’t know when, or if, the price of Dai will fall again
2)Since you cannot complete all the steps of the cycle simultaneously, you are stuck with long exposure to ETH while you wait. You want to sterilize the risk by shorting ETH, but that incurs an additional borrow cost.
3) You have an additional cost of capital for locking up the part of sterilized ETH you didn’t borrow Dai against, which is at least 33% (since you can draw $0.66 Dai for every $1 ETH). The cost of that is the risk-free rate of USD.
4) There is an additional cost for closing the CDP.

Now here's the catch - according to them (and in fact I have no trouble warming to the idea), Maker DAO is not about the stablecoin! The stablecoin in the model is rather a feature, but not the main value proposition. Instead, that is long leverage and treasury/payroll management for ICOs by generating liquidity on the ETH collected in the fundraise, while acting as a swap of sorts, allowing borrowers to defer capital gains tax to a later date. As Hasu and Su Zhu put it, the general misunderstanding is that the purpose of Maker DAO "is not to create a scalable, censorship-resistant stablecoin. It is to be able to generate censorship resistant stability for anyone holding a volatile censorship resistant asset." From that perspective, short term peg fluctuations don't matter as much, as they do for pure stablecoins such as Tether and USDC, so long as long term stability is guaranteed.

Of course, working on a single collateral model, with the collateral being a highly volatile asset such as ETH, is not exactly catering to future guarantees of stability. The relative unpredictability of the stability fee will persist, as long as ETH volatility persists, erroding the overall value proposition. However, let's not forget that single collateral is only v.1, with v.2 being just around the corner - but that's perhaps a story for tomorrow's edition!

Before we close today's session on Maker DAO, I'd like to highlight an important realization that I feel is generalizable accross a slew of smart contract applications and cryptoassets overall; in their current form, they are tools with a set of properties that loosely define the opportunity space they exist in. In other words, their utility is driven by the narrative around them, as much as the narrative around them is driven by their utility.

Considering that we are talking about pieces of technology that are so new they are often framed by the devs themselves as "experiments", and are built on open standards, it shouldn't be surprising at all that their utility is also open to interpretation.

In fact, this is actually a crucial piece in the process of cryptoassets and open protocols to finding product-market fit, and if anything, it's an extremely powerful feature (cc: hivemind). As these protocols mature, the ratio between narrative and utility should skew towards utility, and the feedback cycles should become shorter as the utility space is more clearly defined.

Sidenote: I saw a tweet recently that really hit the spot; the metric that really matters in DeFI, is the total value of assets locked in those smart contracts - not the DAU. If that is true however, then a follow-on point is also true; UX needs to improve beyond just simple interfaces and short user journeys. For these apps to become sticky, they also need to be able to work in a "set it and forget it" way - consider that Maker DAO, as it stands is a high friction application, that requires CDP holders to monitor the stability fee and act accordingly.

Simple interfaces and short user journeys are great for reeling users in; what happens after users are in, is largely what will decide whether or not these users will stick with the platform, once the novelty wears out.

Part 4

After a short intermission, we continue today with our dive into Maker DAO and today we discuss more recent developments, as well as the impeding multi-collateral DAI and its potential implications. The Maker DAO team released the code for Multi-Collateral Dai back in November 2018 and contracts have been deployed to the Kovan testnet (Ethereum). Through that process, the various contracts have been put to the test, undergoing several iterations. A key - and highly attractive - feature of the Maker DAO Mainnet is the improved UI and UX, with a revamped governance page (already in place) and a new CDP Portal.

Now, in various parts of editions 1-3, we talked about the stability of DAI, and how ETH volatility is major pain point in the process of keeping the DAI stable. Multi-collateral DAI will gradually introduce the ability to collateralize pretty much any asset on the Ethereum blockchain. Now at first glance this seems to me as a double edged sword; on the one hand you are opening up to a multitude of assets, increasing the potential collateral pool and thus amplifying the amount of DAI that can be issued, smoothing out some of the demand side pressures in the stability of DAI. On the other hand however, given that we are still some way off from tokenized T-Bills (on the Ethereum blockchain no less), it seems that by widening the spectrum of tokens you are introducing in the collateral pool, the total volatility that the collateral pool is exposed to, increases as well.

Consider the following plot for example. The volatility (sigma) of the daily returns of a swathe of popular ERC-20's that could be used in the collateral pool, in the year between Oct. 2017 and Oct. 2018, was actually higher than Ethereum's in ALL cases. For specificity, ETH's volatility was at ~0.05, while the average volatility of the basket of tokens (weighed by market cap) stood at ~0.11 - almost double of that in ETH.

Let's then take a look at what would happen as to the total volatility exposure of the collateral pool as we gradually introduce more ERC-20's to the mix.

As you can observe volatility exposure increases by more than 50% once 40% of the basket is composed of ERC-20's. If we approximate the likely final distribution by market caps, it would look more like a 15% ERC-20's in the total pool - which would approximate to a volatility profile of 0.68, or a 20% increase from current levels.

In any case, not insignificant, and to compound that it is unclear how much bandwidth the higher absolute value of the collateral will provide in offsetting the increased volatility effect. At least at a high level, if demand grows in line with supply, and the collateralization ratio remains the same that shouldn't impact the net effect of volatility.

There is an argument here to be made about the unavailability of the ERC-20's that are locked in collateral, helping smooth out volatility and put upwards pressure on the price, due to scarcity, but also, I imagine that the prime candidates for inclusion in the pool, are tokens that are already in cold storage and do not actively participate in the market - inc which case, the net effect is still intact.

However, there is another policy lever introduced in Multi-Collateral DAI (MCD), that will likely help control that increase in volatility; the DAI Savings Rate (DSR). A person who holds DAI can lock and unlock DAI into a DSR contract at any time. Once locked into the DSR contract, DAI continuously accrues, based on a global system variable called the DSR. There are no restrictions or fees for using DSR other than the gas required for locking and unlocking. The DSR is funded out of the Stability Fees paid by CDPs. For example, if the average Stability Fees collected on CDPs is 3%, it could be used to fund a DSR of 2%. The purpose of this feature is to help balance supply and demand of DAI and will be one of the monetary policy levers that decentralized Maker Governance can control.

It is a global parameter that needs to be adjusted often to deal with short-term changes in market conditions of the Dai economy. This is in contrast to Risk Governance, which is a long-term process that involves setting Stability Fees, and other risk parameters individually for each collateral type

This is more or less what we know so far. To avoid overload, I'll opt for a semicolon here and defer the final piece of the Maker DAO dive to Monday, where we'll look into the weaknesses of the model and the various criticisms around it. Till then, enjoy the weekend all!

Part 5

We kick off the week with the final piece of the Maker DAO dive, focusing on the criticism around the system and product. We'll attack the points one-by-one, kicking off with the relative unpredictability of the stability fee.

1) Unpredictable stability fee: As we've explored previously, the stability fee is an annualized figure, paid once a CDP is closed. Once the fees have been collected, the smart contract platform transfers the MKR to a contract called the Burner. As no one has the ability to remove funds from that address, all MKR that is contained there is forever removed from circulation. Given that governance is executed by staking MKR, big holders have more control over the stability fee, thus making the EV of participation in the collateral pool more unpredictable for smaller entities - and thus governance more centralized. This might be a feature, as wider participation might just not be possible as it would ramp up the amount of noise in the decision making process. Participating in governance, as the value of the platform increases, comes with a prerequisite of increasingly higher resource commitment (time & effort).

2) Poor off-chain stabilization ability: You can't trade DAI for USD, therefore it becomes difficult for arbitrageurs to participate, as once again the EV is tied to collateral volatility, while the multiple steps necessary to close a CDP introduce a significant amount of friction. Therefore, as on-chain arbitrage is the only way to control stability, Maker DAO faces the same problems as other centralized digital currencies, including counterparty risk, and the risk of getting its accounts frozen or seized and shut down by governments.

3) Maker DAO doesn't scale very well: Given the above, as we have discussed in previous edition, it seems that Maker DAO's ability to scale, is somewhat capped - as its scalability is supply side driven and the responsiveness to demand side movements is somewhat capped by a time-lag. Hasu and Su Zhu did a good job in demonstrating this, so I am borrowing directly from their analysis: "...the combined market cap of all crypto assets listed on coinmarketcap is less than $300 billion. Let’s assume that Maker’s mature version will be able to include CDP types for crypto assets that cover 50% of the crypto market cap, and people want to put 10% of all tokens in CDPs. Then, with a 300% average collateralization they’ll have a maximum Dai market cap of $3.3 billion." Therefore, Maker DAO is more accurately described as a lending facility and less so as a stablecoin issuer.

4) The MKR token could be deemed a security: An often overlooked implication of the incentive structure here, is that the probability that the MKR token is deemed a security is non-trivial. A lot of that will be decided upon the case history that the SEC builds, and the degree to which they think that the pool of participants in the Maker DAO governance is decentralized or not.

5) Low volumes in MKR introduces a liquidity risk: While Maker DAO ranks highly in mcap (~#30), its liquidity profile tells a different story. The lack of trading volume translates into substantial liquidity and systematic risks for MKR holders, as they might incur a substantial loss in transaction value when they liquidate or enter into positions. In some instance, illiquidity could cost investors 20% of the token price, and it is important to note that illiquidity increase during bear markets, which is further exacerbated in a black swan event where MKR is liquidated to cover losses.

6) Short-term multi-collateral DAI: As we have seen in recent communications of the team, the first stage of multi-collateral DAI, will include ERC-20 derivatives of ETH. However, almost all ERC-20s exhibit a higher beta than ETH, implying that stabilization in MCD will be *harder, not easier* when MCD is introduced.

7) General concerns with the role of Oracles in the model: Oracles, in the Maker DAO model, are trusted third parties that provide the system with price feeds. The Maker DAO team, aggregates those feeds through a contracts called the medianizer - more on which can be found here. The system is fairly well thought out, however, be that as ti may, oracles still constitute an attack vector, as they can be corrupted. More on what the team is working on wrt oracles here.

The progress that the Maker DAO ecosystem has been truly remarkable, and in the process it has created millions of USD in value, particularly for its early investors and team. However, as today's section demonstrated, one must be cognizant of the fact that the project is still an early stage foray. A lot remains to be proven, and more certainty in MKR as a potential investment will be enabled once we know more about the MCD upgrade. It appears that at the moment a lot about Maker DAO is in a state of flux, as little is known about the impact of adding ERC-20s to the collateral mix, MKR's legal standing, and the reaction of the marginal CDP holder to price hikes in the stability fee. Hopefully this series provided readers with a good overview of how the model works and the various nuances around it - ultimately building a mental framework, upon which future judgements can be made on.