What’s the Deal with Dai?

 

The MakerDAO and Dai system has recently come under increased attention and scrutiny in the crypto industry. Specifically, critiques follow the recent and frequent increases of the Maker stability fee (interest rate) in response to Dai struggling to hold a 1:1 USD peg to the USD for most of 2019. Maker has now achieved a reasonable peg after raising the stability fee to 19.5% before finally brining DAI back above 0.985USD.

 

Lets explore the mechanisms that lead to this deviation of the DAI peg and see if the criticism of the MakerDAO system is justified.

Snapshot of the MakerDAO

We have looked at Maker previously (RSC Newsletter 10 Feb 2019) so I will only briefly summarise how the MakerDAO and Dai system works here so we can analyse this problem further.

 

Maker is a protocol and DAO operating on the Ethereum blockchain. It comprises a two token system with the first token MKR acting as the governance token and the second token DAI being the product of the system, a collateralised stablecoin. Basically, people can lock collateral tokens in the Maker smart contract and take out a loan called a collateralised debt position (CDP) which is denominated in DAI. Users may borrow up to the value of 50% of the ETH locked up.

 

Currently, only ETH may be used as collateral (single collateral Dai, SCD) and the platform plans to upgrade to allow more tokens to be collateralised (multi-collateral Dai, MCD). Holders of MKR provide governance and vote on setting the stability fee is which is akin to an interest rates for borrowing DAI.

 

The stability fee is what is used to control the supply of Dai entering the system to maintain the 1:1 USD peg. The following graphic details how the peg and stability fee system interrelate.

Figure 1 – Overview of the Stability Fee decision making process by MKR holders (left graphic borrowed form Ray Dalio)

What are people using CDPs for?

At the moment, ETH is the only collateral token available and the reason you would open a CDP today is:

You do not need to sell your ETH collateral so you maintain a net LONG position on ETH

You gain access to additional funds (DAI) which you can sell on a secondary market for another token (going LONG on another asset, usually ETH). This is the most common use case we see today.

Alternatively, you could lock the DAI in a DeFi money market contract like Compound.finance, Dharma or Nuo Network if the interest rate is higher than the stability fee on Maker earning interest rate arbitrage. This is less common as the difference in rates between platforms is relatively small.

Taking out a loan to use the DAI as a payment currency for goods or services, literally a self-funded loan.

 

So the current dominant use case for CDPs right now is to go leveraged LONG on ETH. Keep in mind that if you borrow against your ETH via a CDP:

If the ETH/USD price falls, your CDP gets closer to liquidation and closing part or all of your CDP becomes necessary to avoid this which reduces global Dai supply.

If ETH/USD rises you can actually borrow more DAI as your locked ETH has a higher valuation which increases global Dai supply.

Dai and the peg

Now recently, DAI has struggled to hold a 1:1 peg, trading between 0.95 and 0.97 for most of 2019. This is in contrast to the 2018 Bear market where DAI did an excellent job of holding a 1:1 peg. This makes sense because as ETH price falls, people must close their CDPs to avoid liquidation making Dai more scarce. Now that semi-bullish sentiment has returned to the market, people are seeing “potential gains” in ETH that they are taking on risk by going leveraged long on ETH using CDPs. It’s a self-financed trading loan.

 

It is possible for an investment in ETH to see significant returns, in crypto a 10% gain in a day is not uncommon. So if a trader sees say a 50% rise in ETH/USD over the course of 2019, then an annualised loan at 15% would net a 35% profit and is an attractive trade. Thus, CDPs are opened, Dai supply increases and the peg falls below 1:1. Furthermore, if the trader is correct and ETH does go up, they can continue to borrow DAI against these new gains, increasing their long exposure and even locking up the newly acquired ETH adding to the problem.

 

This is the main reason DAI is struggling to hold a peg of late. It is a sign that the market is generally Bullish on ETH and is willing to accept the cost of finance to go leveraged long.

 

MKR holders have now voted in EIGHT(!) increases to the stability fee over a period of 12 weeks to increase the cost of finance from 0.5% to 19.5% in an attempt reduce the Dai supply by forcing CDPs closed. The raises from 0.5% to 11.5% were largely ineffective and only after the further increases did we seeing signs of CDPs closing and a gradual improvement of the Dai peg.

Figure 2 – Chart showing ETH price (grey), Dai Supply (Yellow) and Stability fee (blue). Note the rapid increase in stability fee from March 2019 to curtail borrowing habits.

There is a clear directional bias here with the Maker naturally geared to be long on ETH except without much incentive to close CDPs due to the expected high return rate on ETH. This could mean stability fees reach and sustain high levels in a bull market, perhaps exceeding 20%+ to maintain a 1:1 peg through the bull market. The rapid increase in fees is likely a lagging reaction to Dai approaching the supply cap of 100M DAI at the same time ETH prices bottomed out and started to reverse.

Interpreting maker platform behaviour

The recent behavior of the Maker platform suggests a few things:

1. During the bear market, Dai was an excellent safe harbor stablecoin – great work
2. The holders of CDPs are currently bullish on ETH and still willing to accept 19.5%+ finance cost to take out leveraged long positions
3. With the stability fee now at 19.5%, some CDPs are being closed. This indicates that some portion of traders see this as a high cost relative to perceived return rate on ETH, an indication they are only mildly bullish.
4. Analysis of the ETH locked in Maker and the DAI supply curve shows when CDP holders transition between leveraging up and closing their positions. This is a reasonable indicator for ETH price strength/weakness and sentiment.
5. The ETH market is net leveraged Long. if ETH/USD sees sell pressure, CDPs will be closed more rapidly, unlocking more ETH supply which could add to sell pressure.

 

Overall, the dominant use case for ETH and Maker today is going leveraged long on ETH which is not ideal but it is indeed a valid use case. We must remember this is early and complex technology. The fact that the Maker system works is incredible and the fact that it works so well is a testament to the development team. With many more features to be rolled out, I expect these issues will be attended to in time.

 

We are watching a brand new and decentralised central bank learning the ropes of monetary policy and economic controls. What is unreal is we can watch all this happen on-chain with the full Maker system open and visible via the platform data portal. If you haven’t had a poke around yet, I would recommend it just to see what is going on under the hood https://mkr.tools/.

is the peg problem a problem?

In a bull market, the cost of finance should be high to curtail rampant and cheap speculation (just like the real economy right…high interest rates when the economy is good…no?…). There will be a competitive market emerge for lending rates between competing DeFi protocols which is starting to resemble a real economy. If Dai does not hold a 1:1 peg this becomes akin to an additional stealth cost of finance as the asset being borrowed has less value. However this does negatively affect the usability of Dai as a stablecoin which is it’s the primary function and must be resolved. As such, it is imperative that a mechanism is developed to incentivize locking up or removing Dai from effective circulation.

 

One potential pathway is to introduce the Dai savings rate which returns half the stability fees paid as interest to Dai holder (the other half continues to burn MKR). I would challenge Maker to eventually have a dynamic savings rate where MKR holders are rewarded for maintaining a good peg. If a Dai peg falls outside say 0.98 to 1.02, more stability fee should go to Dai holders and less to burning MKR. That said, if the peg goes too high, would negative interest rates be enacted to force release of Dai back into the system, potentially yes?

 

Another bold approach is actually to allow a negatively correlated asset like dydx Short-ETH to become collateral in Maker. This would allow traders to leverage in both directions and actually make for an efficient ETH market and behavior as a genuine commodity money. Until negatively correlated assets are available, Maker might suffer this Bull-Bear dichotomy where it only works half the time.

 

I would expect that the longer Maker exists and the more data is collected, the response to such events will become more seamless and appropriately judged in advance. If we assume that Multi-collateral Dai and the Dai savings rate are implemented as planned there is still reasonable to expect the Dai peg to improve under bullish ETH conditions.

Critiques

I do have a few critiques of Maker and how this recent stability fee adjustment has been managed.

 

What I do NOT want to see is more collateral sources based on ERC-20 tokens which are all directionally correlated to ETH. By locking these tokens in CDPs, it also means they are not out there “generating value” in the system they were designed for. To me, this indicates a recognition that we have created internet funny money and Maker might be creating a use-case for useless tokens. Digix Gold (DGX) is a sensible exception as it is a token backed by gold assets and thus has underlying value. Governance tokens like REP, DGD or ZRX should definitely NOT be available as collateral as it locks away tokens which should productive members of ETH-society. Perhaps Security tokens or index tokens will fill this brief eventually.

 

Only 2 weeks were provided between each stability fee vote was put to the DAO governance. More time is required to allow the system to reach equilibrium after each increase to properly assess actual performance. I suspect this is a symptom of being the first significant market reversal Maker has experienced and is a teething issue. The stability fee was at 0.5% before this kicked off which is extremely low and likely due to the desire to onboard Dai users in the bootstrapping phase.

 

Another major issue is that the Maker model is eventually supposed to mint MKR tokens if the peg runs too far away, making MKR holders the lenders of last resort. This feature is currently not active and thus MKR holders are increasing the stability fee and burning more MKR yet with no risk to their token value. This is a somewhat rigged system and it is actually the CDP holders who take on this risk instead. In the event of a Dai price crash, CDP holders will get back less ETH than they put in. The update to MKR token functionality needs to be rolled out to fix this.

 

My final gripe is the long tail distribution of CDP holders. Right now, over 80% of CDPs are held by the top 100 CDPs who are uber-rich ETH holders with an average of 16,886 ETH locked up (that’s $2.53m at a price of $150 / ETH). These CDPs remain largely unchanged in the last month comparing the data from April (blue) and May (black). The total CDP value owned by the top 100 actually increased during this time indicating that it is small accounts who are most impacted by these stability fee increases.

 

Maker has a cap on how much Dai can be taken out by the system, currently set at 100M Dai. If ETH continues to appreciate, these rich folk can continue to borrow against their holdings which may continue to restrict new people from accessing the Maker service and drive the stability fee up higher. This also created a massive leveraged Long ETH position in the hands of a few hundred accounts.

Figure 3 – Analysis of the top 100 CDP positions showing a heavy concentration of ETH held by a handful of accounts. Also note almost no change over April-May comparing blue to black.

Last thoughts

Maker is easily the most exciting token on the market in my opinion. It’s the only one I see representing an actual valid use case for “utility” tokens – that is for decentralised governance and shares in a digitally native organization. This is something that was not possible before blockchains. The Maker developers have created a world first and world class piece of technology with the potential to disrupt finance and banking.

 

However, I am keen to see the model expand beyond a one directional market for ETH and a sounder use case for ETH needs to be developed than going LONG on ETH. Right now, the DeFi ecosystem and the Ethereum 2.0 security relies heavily on ETH token price appreciation and is very much an internal economy. Hopefully, a survival toolkit will emerge for ETH and DAI to protect their value in the event of price shocks as well as some increased external use cases aside from speculation. MKR token holders need more incentives to anticipate the market as well accept more exposure to the risk of their product.

 

It is important to note is the general outcry of people claiming 19.5% interest rates are too high is indicative of how desensitized we have become too low interest rate policy. I believe that during the good times, interest rates should be high to prevent unchecked debt levels. This is something we do not see in the modern economy where central banks exist on near zero interest rates and practically free debt for decades. This invariably ends in a correction and one which is always much larger the longer unchecked leverage persists.