Report: Collateralized Stable Coin

Introduction: What are Collateralized Stablecoins 

A collateralized stablecoin is a token, typically implemented with a fungible token standard on a blockchain supporting smart contracts, whose value proposition is to remain pegged to a real world currency. In most cases such currency will be the US dollar. What distinguishes collateralized stablecoins from other classes of stablecoins is the mechanism with which they are able to maintain their value stable. In particular, collateralized stablecoins derive their value from a certain amount of collateral which is locked in a smart contract and serves to peg a certain amount of outstanding stablecoins. Most of the time the collateral is represented by a volatile asset, therefore, one of the major challenges for the projects present in the sector is to find the right balance between capital efficiency and a healthy collateralization ratio. 

In order to analyse this peculiar class of crypto assets we are going to use a general framework looking at the sector from three different perspectives. In market design we are going to look at the environment in which the stablecoins are traded. In mechanism design we are going to analyse the set of algorithmic rules governing the stabilization methods behind the token. Lastly, in token design we are going to look at how a secondary token could be used to improve the functionality of the system and provide a means to allocate governance powers.

Glossary 

  • 2nd Order Stability Index: measures how the stable coin price returns to stability looking at the 3rd and 4th motion of price returns
  • Circulating Supply: the number of tokens that are publicly available and circulating in the market.
  • Density metric: the ratio of the number of edges to the maximum possible edges
  • Eigenvector centrality: a measure of the influence of a node in a network
  • Hard peg: A peg that is hard coded in relation to other assets (eg. sUSD being hard coded to 1.00)
  • Lorenz Curve: a measure of inequality of wealth distribution within a system
  • Market design: the design parameter of the ecosystem that the participants transact in and the token exists in. 
  • Mechanism design: reverse game theory or rules of the game. Instead of participants choosing a strategy based on the outcome of their choices, mechanism design is the rules of the game, to constrain the behaviours of participants, towards a desired outcome. 
  • Soft peg: A peg that attempts to maintain 1 but not in relation to pricing other assets (eg. DAI tracking the price of ETH dynamically)
  • Stability Index: measures how quickly the protocol recovers to the acceptable price range.
  • Token design: design of the token in the ecosystem.

Metrics Discussion

General Framework 

The general Economics Design framework is used to analyse the fundamental economics between the various incentive mechanisms and token design. It looks at the aspects of market design (digital environment), mechanism design (structural rules of the environment) and token design (mechanism and incentive policies of the token). 

We analysed the various possible metrics within the framework and shortlisted the relevant ones. These metrics are deemed the most important as they affect the robustness of the design for such collateralised stablecoins. 

The research is limited to onchain data from 1 January 2021 to 24 May 2021.This is to ensure that the time period used is consistent and it removes any market shocks in different time periods. 

Market Design 

Market design is the ecosystem that the participants transact in and the token exists in. It is important to constrain the digital environment because market design defines the structural limit of the mechanism’s effectiveness. 

[Market Design] Introduction to quantitative metrics

As discussed above, market design is the digital environment that the tokens exist in. That exists beyond the protocol itself, and also includes the acceptability and usability in other protocols and ecosystems. Thus, in the qualitative metrics, we analyse the token with respect to the community holding and using it. 

[Market Design] Quantitative Metrics

To quantify the level of adoption of the network, we took the transactional amount and volume between all users in the specific protocol in the specific period we defined above. Being a stablecoin with the purpose of day-to-day exchange, the market level analysis and measurement are very important to quantify the usability of the stablecoin. That is because it is the main objective of the token’s existence. This inherently defines the “demand side” function of the stablecoin.  

Demand side function is to quantify the onchain utility of the stable coin.

Benchmark metrics: The metrics will only make sense when compared against a general benchmark. To create the benchmark, all 6 protocols are added together, acting like one giant network. The same metrics are applied, thus giving us a benchmark for the overall network to compare against. 

In the last metric of distribution coefficient, another benchmark is used. That is the distribution of USDT as a control variable and the average of all 6 distribution coefficients. USDT is used as it is the most liquid stable coin in the market, hence the benchmark token. 

Note: As some protocols only started after 1 January, we gathered onchain data from the time they started, till 24 May 2021.

[Market Design] Density of Macro Market

The lower, the better

One measure of market decentralisation is the density of markets. It is possible to have metrics supporting a robust market design, but if the network is centralised around a small percentage of users with high social capital, that suggests that the market is not robust enough. This is calculated by looking at how widely-used the token is, when looking at the transactional patterns over the period of research.

To compare the use of metrics, each density metric of a protocol is compared against the overall average density. The higher the figure, the more decentralised. That suggests that each individual is more equally connected to another individual, instead of just having an individual connecting to most of the network. 

LUSD is the most dense and DAI is the least. This metric alone suggests that on average, each wallet is more connected to another wallet that holds LUSD in the system than wallets holding DAI and being connected to other wallets holding DAI. 

[Market Design] Centralisation Degree of Wallet Addresses   

The lower the better

Unlike the density of the market that focuses more on the overall network, it is possible to measure the centralisation from an individual user. This measurement quantifies the connectivity of a wallet address by understanding the amount of transaction that interacts with this address. This also includes addresses of protocols, not just individual addresses. The lower the degree of centrality, the less “influence” the wallet has, the more decentralised the network. 

Most stablecoins have a higher degree of centrality than the benchmark, combining all transaction interaction within an ecosystem. Whilst this could suggest that the tokens are more centralised on average, it is also mindful to realise that most of the tokens are still relatively new. 

DAI is the most decentralised, in these metrics and LUSD is the least. Relevant to these metrics, the top addresses that are the main connector to all other users are exchange tokens. This is predictable as exchanges are the one of the main methods to use the stablecoins, be it for arbitrage, purchasing of other tokens or providing liquidity due to rewards for liquidity providers. They also hold the largest amount of tokens by percentage and exist to facilitate transactions between users. Hence, they are also responsible for the high degree of centrality. 

For stablecoin protocols, the more gateways of transactions, the better. For example, DEX like Uniswap, Sushiswap and Curve, and CEX like Binance. This is likely because these addresses act as facilitators to enable transactions between users.

[Market Design] Eigenvector Centrality

The higher the better

The last metric in defining the market network of the token is eigenvector centrality. The focus of this metric is to understand the secondary effect of influence for a wallet address (user). This metric quantifies the level of influence that the user has and the quality of the network that the user has. In aggregate, this gives us a better picture of the overall network’s quality of (de)centralisation. The mechanism and token design is not just an analysis of primary effects, but also secondary and tertiary influence on user’s (network’s) behaviours. Thus, this metric can quantify them in the aspect of stablecoins. 

Eigenvector centrality provides a more robust understanding of the network, because it puts into perspective the two decentralisation metrics discussed prior. A higher score suggests high quality of centralisation. That means the user, on average, transacts with another user of high scores, which suggests a high secondary influence in the network. The average score in the network is calculated to define the average score of the network. 

This can be used to understand the market by quantifying the speed of implementation amongst user behaviours when mechanism changes. Onchain transaction is the behaviour to analyse, to understand if the mechanism and token are well designed.

LUSD has almost twice the score compared to the combined benchmark. This suggests that on average, the wallets in LUSD connect to wallets with high transaction volumes. This suggests that LUSD is transacted and used more than the combined average. On the other hand, DAI has the lowest eigenvalue centrality that suggests it is not used as much in comparison and has lower secondary influence.  

[Market Design] Change in Circulating Supply

The higher the better

ALUSD

SUSD

LUSD

VAI

DAI

FEI

101.40%

12.74%

353.53%

-40.34%

32.22%

-59.26%

Due to the fact that change in circulating supply means more collaterals are added into the system, a positive growth in circulating supply suggests a healthy sign for the ecosystem. The higher the growth, the higher the demand by investors. 

LUSD and ALUSD are seeing a huge positive change in circulating supply, but this is mainly due to the fact that the protocol just started then, hence the large positive change. However, also starting at the same time, FEI and VAI are seeing negative growth. DAI and SUSD are in a more steady state, with a lower positive growth rate. 

[Market Design] Change in Transactions 

The higher the better

ALUSD

SUSD

LUSD

VAI

DAI

FEI

-62.98%

120.03%

-94%

110.76%

71.11%

-18.26%

The change in daily transactions volume tells us the usability of the token. The higher the value, the more people are transacting in this token. This means that more people are accepting the token, or that the token is used by more protocols. In any case, this spells good market design of the protocol. 

SUSD and VAI have high positive changes, which suggests usability of the token, followed by DAI. ALUSD, LUSD and FEI are showing negative changes. This could be due to the protocols just launching, and the utility is still limited to just purchasing the tokens. On the other hand, DAI and SUSD are used in many other protocols, which increases the usability. 

[Market Design] Token Velocity: Transfer Per Circulating Supply

The higher the better

The transfer per circulating supply shows the empirical velocity of the token, which provides us with the insight of its usability. A higher figure suggests a higher velocity, which makes a more robust ecosystem.

To calculate the network growth, a good measure is to compare the number of transactions (usability of stablecoin) against the total circulating supply. In general, the higher the amount, the better.

A higher percentage shows the increased usability, thus a value of the stablecoin. For newly launched stablecoins, a general downward sloping trend is not uncommon. We see that with alUSD and LUSD. As it matures, it should stabilise around a certain range. We see that with DAI and SUSD. The higher percentage for DAI  in the first quarter of the year can be attributed to a considerable price surge in Maker’s MKR token, which could have driven up the volume of transfers. FEI is almost at negligible growth, potentially due to the massive circulating supply in the network. VAI, on the other hand, sees an upward trend. 

Situations that can skew the percentage are the launch of a token and new announcements. In the example of Maker, it could be due to the financial release or an executive vote on the governance² to increase the spending and surplus amount in Maker. 

To accurately compare the fluctuations of the stablecoins, the stablecoins are rescaled to the same x-axis. Generally, SUSD, FEI³ and LUSD are stable with mild fluctuations. DAI has a downward trend, plummeting from January and stabilises from March. The percentage for DAO shot up on 24 April, potentially due to the announcement of an upgrade on the protocol’s liquidation system. VAI faces more volatility with an upward trend, peaking twice in May. The rising trend is attributable to a sudden price increase that causes massive liquidations, which in turn fueled an increase in the volume of transactions.

³These are prices on exchanges, not the real prices of FEI after the direct disincentives in place. 

[Market Design] Growth of Usability

To quantify the trends of usability, the growth rate of daily token transfer is analysed. ALUSD, DAI and SUSD see a stable growth rate. VAI, FEI and LUSD see an upward trend in their growth. For new protocols and new tokens, an increasing growth of daily transfer is expected while protocols that exist for a longer time should see a slow growth trend. FEI and LUSD see the highest growth trends. 

[Market Design] Distribution coefficient of transactions

The lower the better

The last measure of market thickness is the distribution of transactions. Again, this is relevant to the application of stablecoins to understand its usability in the system as a whole. This measurement quantifies the distribution of transactions using the Lorenz Curve. Instead of commenting on inequality, we are measuring the spread of transactional volume to compare against other protocols discussed. 

The distribution is ranked from 0 to 1. The lower the number, the more “equal” the transaction amount, suggesting a wider use-case. For example, using the stablecoin to purchase things. A higher number suggests larger volume movement, like high frequency trading, arbitrage, and rebalancing of vaults via the protocol. The detailed discrepancies are based on the specific mechanism design of the protocol.

To make sense of the data, the benchmark is USDT, the most liquid stablecoin of the period. USDT has a high coefficient of 0.9280. This is reasonable as being the most liquid asset, the large valued transactions are in USDT, increasing the coefficient. 

Ranking the 6 protocols, ALUSD has the best distribution coefficient while FEI has the worst. This is likely because FEI has a mechanism called PCV that skews the distribution of transaction volume closer to 1. ALUSD, being the only token collateralised using another stablecoin, is likely to see a more even distribution as it is least affected by fluctuation in collateral asset prices. 

Market Design Conclusion

In general, LUSD and DAI stood out the most in the market design analysis. LUSD is high in density, low in degree centralisation and high in eigenvector centrality. It has a general stable growth rate, albeit in the downward trend, for token transfers. Lastly, it has the lowest distribution coefficient of transactions. 

That suggests that LUSD’s users are most connected to each other, while having influential users that are connected to other influential users. Based on the transactions, it is declining at a steady rate, as it is a relatively new protocol, compared to other new protocols like ALUSD and FEI which are declining at a quicker rate. The distribution of transactions are also most decentralised in LUSD, compared to the other 5 stablecoins.

On the other hand, DAI has a low density, high degree centralisation, high eigenvector centrality, stable growth of transfer and concentrated distribution of token transfer. This suggests that the market is widely distributed with transactions more concentrated in a segment of users. The growth rate is also flatter. This is not surprising, considering DAI is the most mature of all onchain stablecoins. The network is likely to be the most distributed, which disproportionately skews the centrality metrics. However, the data is aligned with the expectations of how such a mature network of a more mature stablecoin like DAI should behave. 

 

Get 20% OFF On Our Book

USE CODE “REPORT”

Mechanism Design 

Simply put, mechanism design comprises the rules of the game, which participants have to follow. 

[Mechanism Design] Quantitative Metrics

Specific to stable coin mechanisms, the mechanism design focuses more on the methods in which the token policy can be changed. This includes governance, making changes to the token monetary policy and its current stability mechanism. 

[Mechanism Design] Through What Decision Making Process Are Protocol Elements/Parameters Changed? 

This crucial metric falls under the broader umbrella of governance. Who can change the rules or the parameters (e.g. interest rates) and how. Most of the protocols have endorsed an open and transparent policy around the evolution of their projects. Communities are incentivized to participate in distributed governance. Which is for the vast majority of the protocols under analysis, the tool for introducing changes in the protocols themselves. The guiding principle behind the scoring is the extent to which many stakeholders are included in the governance.

MakerDAO scored the highest, which is understandable as it is probably the most successful DAO experiment to date and has had the longest history, compared to the other protocols, of successfully responding to issues. sUSD with its elected council also ranked high. It is most reflective of a democratic system we know from the off-chain world.

[Mechanism Design] How many IPs passed in the ecosystem?

The number of improvement proposals passed in the ecosystem gives an idea both of the effectiveness of the governance system and the liveness of the development process. When relevant, the number of proposed versus passed IPs can provide an insight into the degree of consensus among stakeholders and the activeness of the decentralised community.

As we can see some protocols choose to adopt a minimal (sometimes zero) governance approach. While this is not necessarily bad and some protocols are very young, no matter how good the design is, certain parameters or even whole elements will inevitably have to change in response to technological development, market conditions, user behavior, competition, etc. sUSD is the highest in this metric, based on the sheer number of passed IPs. This might be misleading since Synthetix is much more than just a stablecoin and thus needs more governance than the others. The IPs include the protocols as a whole, not just sUSD itself. Its representative system also allows for faster decision making. Maker needs to be commended for passing this many IPs with more stakeholders involved. 

[Mechanism Design] What happens when price is below $1

When the price goes below $1, the token loses some of its functionality, for example users holding the token as a store of value lose some of their wealth. The stablecoin is not, in fact, stable. The remedy needed to raise the price back to one dollar must either aim to increase the demand or decrease the supply, or both.

[Mechanism Design] What happens when price is above $1

Similar to the previous metric, the stablecoin loses its stability. Since the solution consists in either raising the supply or decreasing the demand the problem is easier to solve than the previous one. This metric tells us all the proposed solutions to restore the peg back to $1 when the price rises above the peg.


[Mechanism Design] Stabilisation method

This metric serves as a thorough description of the arbitrage in place to keep the price stable and as close as possible to $1. The solutions are varied and involve both soft and hard mechanisms that induce agents to participate in the protocol, facilitating its stabilization.

Providing commentary on stabilization methods is a difficult task, as it’s pretty much impossible to make judgements without empirical support. Thankfully we go into details on the observed stability of the protocols’ pegs in other sections of the report. It is important to understand the level of intricacy the protocols must go through to achieve relative stability. 

We gave sUSD the highest score, as it integrates stabilization into a broader ecosystem where the stablecoin is an essential moving part and judging by the project’s success we can say the stablecoin is doing well. Maker and Liquity are not far behind and their stabilization mechanisms are much more intricate. This can be both a pro and a con, especially as crypto gains mainstream adoption and users with less technical knowledge start using these protocols. There is certainly an economies-of-scale aspect with stability and Maker has proven its model time and time again, becoming one of the largest coins on the market. As stability is certainly the most important feature of any stablecoin, we will speak further on this topic in other parts of the report, connecting it to empirical data. You can find it under Token Design. 

[Mechanism Design] Peg Verification Methods/Oracles Used: How Is The $1 Peg Verified ?

In order to verify that the token is pegged to $1 a verification method is needed, so the protocol can trigger stabilization events. Practically, there is a need to obtain information from a third reliable source: oracles. When talking about price oracles can take different forms. In particular, besides normal oracles, there are Decentralized Exchanges APIs providing price data. The different solutions for peg verification, which, as we will see, are heavily skewed towards decentralized oracles and DEXes.

Lots has been written on the topic of oracles (and their problems) so we won’t go into detail here. We need to, however, understand what makes an oracle a good fit for a stablecoin. The consensus seems to be that, the more diversified the information the better. As few stablecoins have liquid markets on CEXes, DEX oracles do a good job, if the markets are big enough. Choosing just a single Uniswap market is risky, for example. Information quality is heavily tied to liquidity and low transaction costs, which might not always be available. 

The protocols that use proprietary oracles are placing a lot of trust in rather centralized structures and the community is likely to move away from such solutions over time. sUSD again has a unique approach where the price of 1 sUSD is fixed to $1 and everything else around it is floating. It therefore doesn’t even need an oracle. Such solutions might be valuable if trust in oracles is eroded by hacks for example. While this is unlikely, they are often called the weakest link for a reason. He who controls the information controls the protocol and governance might not have the time to react effectively.

Token Design

Token design is the rules embedded within the token that the token has to follow.

[Token Design] Quantitative Metrics

Specific to stable coin mechanisms, the token design includes the monetary policy of the token, secondary token in the ecosystem and empirical evidence on how successful the design of the protocol has been.  

[Token Design] What Are The Incentives To Various Agents In The Protocol

This metric focuses on which economic benefits are offered to an agent for her to engage with the protocol. We look for tangible links between the incentives embedded on the protocol and the actual result shown by data. These incentives help any type of user receive a return and thus build a robust ecosystem

Maker DAO scored the highest among the protocols present, the reason being that the MKR token is placed at the center of a cleverly designed incentive system. The system combines returns from fees, governance and MKR issuance to align participants’ incentives. Borrowers and DAI holders use the protocol following the rules set by MKR holders. These agents do not have much responsibility and pay a fee in exchange for a service. MKR holders on the other hand have the responsibility to decide on key protocol’s parameters, for this reason they receive revenues in the form of transaction costs. On the other hand, when debt positions become insolvent, the losses are covered by auctioning newly issued MKR’s tokens, such a mechanism punished pre existing MKR holders with dilution and constitutes an incentive to maintain good governance. 

[Token Design] General Statistical Measures

The price stability of tokens can be determined by statistical measures such as standard deviation, skewness and kurtosis.

Standard Deviation indicates market volatility, measuring how widely prices are dispersed from the peg. If prices swing widely up and down, the standard deviation returns a higher value and indicates high volatility. VAI has the highest volatility, with prices falling below the peg the majority of the time. Another token with high volatility is DAI. Although DAI features a normal distribution, it is more volatile compared to USDC which suggests wider swings in prices from the $1 peg. ALUSD, FEI, LUSD and SUSD are relatively stable with low volatility. 

Skewness measures the asymmetry of distribution. Perfectly symmetrical data have a skewness of close to 0. Most of the tokens (ALUSD, DAI, FEI and VAI) have a distribution that is highly skewed, with skewness values greater than 1. VAI is extremely skewed at 2.45, suggesting the token price often falls below the peg. SUSD token price might drop off the peg occasionally as the distribution is moderately skewed. LUSD is the most volatility-free, with its distribution being approximately symmetric.

Kurtosis identifies extreme values of a distribution. In general, the higher the kurtosis, the more likely the future price of the token will encounter extreme deviations from peg. ALUSD, DAI, FEI and VAI have a higher kurtosis value compared to USDC and might have a high risk of deviation. This is especially true for ALUSD and DAI, which approximate the distribution of USDC. It is explainable for VAI to have a high Kurtosis, given the extreme skewness. 

In general, ALUSD is the most stable out of the 6 tokens. DAI may have a similar distribution that looks approximately normal, but the token is highly volatile, with a higher risk of attaining extremely lower or higher prices. 

[Token Design] Percentage at, below and over pegged price

To quantify if a token is underpegged or overpegged, we looked at the percentage below, at and above the pegged price.

Majority of the tokens are overpegged, with the token price rising above the pegged amount of $1. Out of them, SUSD have a staggering 92%. FEI and VAI are under pegged, with VAI having a percentage of 99.92% falling below the peg.

Surprisingly, USDC is frequently overpegged, at the price of $1.01. If we analyse $1.01 as the peg price, both ALUSD and DAI will be considered relatively less volatile. As when the peg is $1,SUSD and LUSD are overpegged while FEI and VAI are still being underpegged.

[Token Design] First Order Stability Index

The lower the better

Feedback loops are important in such closed ecosystems. As such, we seek to measure it. Specifically, we created the First Order Stability Index, which is the normalised feedback loop index, compared across the tokens analysed. In this case, the lower the feedback loop index, the better. 

Stability Index

ALUSD

DAI

FEI

LUSD

SUSD

VAI

Feedback Loop Index 

7

2.2

5.63

2.6

3.95

11.15

The Stability index measures how quickly the protocol recovers to the acceptable price range. The index considers the days spent within and out of the acceptable price range for us to analyse how stable the prices are. It is normalized to 10, as to fit on the 0-10 Radar chart. 

The fundamental quality of any index is taking different data sources and combining them into a single-scale data point that can easily be compared across samples. We looked at daily price data for 2021 (the actual time frame differs across protocols, as some are younger than others, but given there were no black swans or significant market crashes during the periods we take this as an acceptable simplification). We set a ±3% (from 0.97 to 1.03) “acceptable boundary” within which we deemed fluctuations to be acceptable and due to no issue in token design. Any time this boundary was breached, we counted a breach and then measured the average time spent in breach of the boundary, in either direction. Once the peg was restored and the price was back inside the boundary, we recorded a “Breach end”. The combination of N of breaches and average time spent in breach forms our “Stability index”

DAI and LUSD are ranked the highest and VAI is ranked the lowest. VAI even came into the negative range, as it has spent considerably longer out of range on average. 

[Token Design] Second Order Stability Index

Stablecoin tokens are pegged to a currency that attempts to offer price stability backed by a reserved asset. While stability is achieved easier in the long time frames along with mean reversion to price $1, in the short run however, a much higher degree of price fluctuation can occur before the price stability measure kicks in. To assess stablecoin price stability at a micro structure level, the “Second Order Stability Index” was developed and tested.

The index jointly assesses the skewness and kurtosis, the 3rd and 4th differential of the price returns of the stable coin(s) price series, to quantify sudden spikes. The kurtosis of price returns is sensitive to large (high amplitude) and fast (sudden non-stationarity) alterations in the price. Thus this is an important price stability quantification metric. 

With probable large deviations, the skewness of the price returns also becomes important to assess and quantify the propensity of the price return to mean (1$ in case of stable coin). For example, in the event of a hypothetical 50% price deviation, the downward scenario starts recovery from price 0.5$ (i.e. = 1$ – 50% of 1$) and needs a 100% upward price movement to price parity of 1$. On the other hand, an initial upward 50% fluctuation will lead to recovery from price 1.5$ (i.e. = 1$ + 50% of 1$), but only require a -33% downward correction to return to price parity of 1$. Thus, in theory, upward high amplitude price deviations are much easier to correct by supply/demand changes (i.e. by fair & transparent market forces) to the stable coin supply than equivalent amplitude downward fluctuation. 

As skewness and kurtosis go hand in hand, a normalised Jarque-Bera (JB) test static was normalised to generate an index between 0 and 1. This index readily allows comparisons between different stable coins as well as comparing the normalised values with disparate and complete different normalised features of other coins. This is also suitable for extending the comparison to new coins as well as new timepans in the future if needed. The following JB test statistic was adapted for the “Second Order Stability index”:

 

ALUSD

DAI

FEI

LUSD

SUSD

VAI

2nd Order Stability Index

0.95

1.00

0.91

0

0.63

0.94

DAI ranked the best while LUSD ranked the lowest. ALUSD, FEI and, surprisingly VAI achieved a raw score close to DAI. The surprise of VAI ranking close to the best (DAI) while having maximum deviation from 1$ peg can be explained by it’s gradual change and consistent price away from the 1$ peg. The lack of high amplitude tail events specifically for price recovery may be attributed to the lack of/inefficient peg mechanism.

Images: Price against frequency of the 6 protocols analysed, with USDC as the control token. 

A good measure to determine price stability is to look at the frequency of tokens at each price level. Generally, tokens with a good price stability will follow a normal distribution. As USDC is the one of the largest stablecoins with negligible fluctuations in price, it will be used as a benchmark in the comparison. ALUSD and DAI are the most stable to peg with a curve following a similar distribution as USDC, with LUSD and SUSD being negatively skewed. FEI and VAI have extreme positive skewness which suggest the price of the tokens often falls below the peg. 

[Token Design] Liquidation mechanisms (E.g. falls below c-ratio)

This metric explains qualitatively the choices made by the projects to liquidate under collateralized positions. We will look through the best solutions in regards to efficiency, safety and implementation

Among the protocols taken under exam, Liquity scored the highest in terms of its liquidation mechanism. By making a distinction between normal and distressed periods the protocol can achieve both safety and efficiency. In Normal Mode the collateralization ratio is set at 110%, a fairly low ratio promoting efficiency. When the protocol goes into Recovery Mode, i.e. when the total collateralization ratio of the protocol falls below 150%, any position with a c-ratio below the total c-ration can be liquidated.

[Token Design] Insolvency solution

At times, the protocols undergo substantial pressures from market conditions and external factors. We try to look at what mechanisms are in place to overcome such pressures or insolvency states.

The newly proposed solution proposed by Liquity recorded a high interest from the analysts of this research. The redistribution of debt if done efficiently should guarantee enough safety for all the participants. Similar score is registered by MakerDAO where its auctions have been proven broadly efficient, except in some infamous cases where extreme moves of the underlying seriously jeopardized the mechanism. 

[Token Design] Secondary token: What does the token take from the protocol?

The vast majority of collateralized stablecoin protocols include a secondary token that incorporates several functions such as governance and value accrual from the economic benefits unleashed by the business model and many others. In this metric we analyse all the fees and economic benefits accrued to the secondary token holders. 

Liquity and MakerDAO scored the highest in this metric. Both the protocols adopt what seems to be a pretty common solution: the secondary token entitles its holders to receive income generated by the protocols fees. 

A natural difference between MakerDAO and Liquity arises from the different approaches the two protocols have toward governance. While Liquity is a rigid system of smart contracts and does not have a governance system in place, Maker DAO puts the secondary token (MKR) at the center of its governance system. In MakerDAO, revenues from the protocol operation come together with the responsibility of maintaining good governance.

[Token Design] Secondary token: What does the token give to the users?

On a similar but inverse note compared to the metric above, here we look at all the auxiliary functions carried out by the secondary token in regards to supporting the protocol in difficult times.  

With the MKR token being at the center of an incentive system aimed at achieving good governance of the protocol, Maker DAO obtained the highest score. MKR is the last resort solution to obtain funds when a debt position becomes insolvent and at the same time provides an incentive to make good governance choices. In fact, if a debt position becomes insolvent, the capital needed to cover the losses is raised by minting new MKR tokens. This way MKR holders are incentivized to maintain healthy collateralization ratios not to incur losses by dilution.

[Token Design] Returns on the stable coin in the protocol

This metric shows the extra return that stablecoins offer to its holders along with the benefits of using the token cross-protocols.

Returns on stablecoins can be obtained in a myriad of different ways in DeFi. However, sometimes the stablecoin protocol itself offers a way to earn risk free interest rates on the stablecoin’s deposits. This mechanism is typically a way to influence the demand for the stablecoin by offering a competitive income opportunity. Maker DAO with the DAI Saving rate smart contract offers a saving rate to users willing to lock their DAIs. The Saving Rate is determined by the Maker DAO governance.

[Token Design] How to calculate borrowing rate & who determines that

This is possibly one of the structural parts of the mechanism where protocol differs the most among them. Different solutions have been tested, and we appreciate most of them. Higher score is recorded from Alchemix that eliminates borrowing fees.

Discussion 

The goal of this research is to understand which parts of the different protocol designs have the most effect on the final scenarios we are seeing play out in the reserve based stablecoin market.

Specifically, the goal of a stablecoin is to optimise for stability. To do that, we divided the discussions into 6 categories: network effects, demand, governance, stability, incentives and secondary token.

Network Effects

A decentralized network spreads transactional information across multiple channels rather than relying on a single point of contact. Rapidly and effectively disseminated information makes the network more decentralized and influential. LUSD appears to have scored the highest across the three metrics (Density of Macro Market, Centralisation Degree of Wallet Addresses and Eigenvector Centrality), while DAI appears to have scored the lowest. We can conclude that LUSD users are more influential and spread transactional information faster than DAI users.

The number of links held between each transaction determines the degree of centrality score. As a result, the higher the degree of centrality, the more popular the protocols are in comparison to the others. With LUSD having the highest degree of centrality, this simply means that people may have used limited or common protocols for the transaction. Whereas DAI, which has the lowest degree of centrality, could indicate that there are many protocols through which people can make transactions, making it more decentralized in this case.

Having a large number of transactions, however, does not determine the network’s influence. As observed, the network scoring low is primarily due to the network as a whole not being well connected to the rest of the DeFi ecosystem. One reason for the observation could be because of VAI. VAI is not a part of the Ethereum ecosystem, but rather of Binance Smart Chain. The exchanges that users use for transactions would be different. As a result, it causes the overall network to be further spread out and decentralized.

Although DAI has a higher degree of transaction decentralization, the network is not as influential, as the transactions may be circling around the same group of users.

Demand

Demand of a stablecoin is determined by the number of transfers over the amount in circulation. Generally, a higher percentage will suggest a higher usability rate of a stablecoin. Ideally, we want a stablecoin with a high and growing monthly transaction, and transaction growing faster than supply expanding.

Stablecoins with high velocity are considered less volatile and hence more stable, making them a suitable store in value, and encourages use in everyday’s transactions. It makes sense that fluctuations in demand would cause fluctuations in price, as the intersection between demand and supply would constantly shift. 

Stability via Velocity 

The token supply in circulation should increase over time, but ideally at a slower rate than the number of transfers. Sequentially, the percentage of demand over supply will increase stability, as circulating supply always exceeds the number of transfers. This helps to increase the velocity of tokens, which is a relevant metrics for stablecoins. 

Relationship between demand, supply and stability 

DAI has the highest usability rate, four times higher as compared to other stablecoins. It is then followed by VAI, which has twice the demand over supply values. However, VAI is less stable with a downwards trend, potentially due its more recent launch. In terms of stability, SUSD, FEI and LUSD are stable with mild fluctuations.

With a higher degree of centrality and lower volatility, LUSD is the best in comparison to the other stablecoins, as transactors are using limited or common protocols for their transactions and the protocol has a stable supply-demand relationship.

Governance

Governance can be divided into hard governance and soft governance. Hard governance includes the rules, mechanisms and ecosystem policy fully embedded in code. Soft governance is where parameters of the ecosystem can be changed by a governing body, like changing interest rates, inflation rates or redemption rates. Governance is not only about decision making, but also resolving any conflicts in the system. Most improvement proposals (IPs) are suggested as part of the resolution of conflicts or flaws in the system, as opposed to active improvement to increase the robustness of the protocol. 

Governance style

In the protocols discussed, Liquity has zero soft governance as it focuses on hard governance. The system dynamically balances the one-off borrowing rate and everything else is fixed in the mechanism. It is questionable how resilient such a rigid system will be. Maker has the most active governance, arguably because they have existed for the longest time. The system is active with frequent discussion of improvements and it is not uncommon to see an improvement proposal broken into segments for approval, improvement or development in the smart contract. SNX is also a mature protocol with an active governance culture, followed by Alchemix, FEI and VAI in terms of governance development. 

Governance capabilities

DAI has the most amount of proposals, discussions and approved proposals. Some proposals are even broken into smaller chunks for further discussion. Of course, it is the largest decentralised stablecoin in the ecosystem and the oldest, thus it is not unrealistic to see the immense community participation. The users are also allowed to change a good amount of economics parameters, which is aligned with the decentralisation philosophy.  

sUSD (SNX) has many IPs passed through its elected Spartan council. sUSD has a fixed value of $1 and it does not change based on any oracle value. Instead, the proposals are mainly on improving the ecosystem as a whole, which allows the increased usability of sUSD as the medium of exchange between assets, as sUSD is just a small part of the SNX ecosystem.

As of the time of report analysis, Alchemix currently uses a multi-sig wallet to make decisions. This is necessary in the start of protocols since it will be difficult to coordinate amongst users. It includes 4 in the core team and 3 from the community to approve of votes. In this model, the ability to change protocol parameters is limited, and the amount of proposals from the community is even less. 

FEI’s 2 votes so far were used to solve the then-issue of the token missing the peg and returning ETH back to users. A handful of proposals are suggested to help restart the faith in the community. All of the proposals are 99.9% yes to 0.1% no, which suggests either a clear communication by the community or voter’s apathy from those who disagree. 

VAI is the last token in the list. Although VAI and DAI use the same model, the quality of governance is not only based on the mechanism for votes, but the quantity, frequency and approval of votes. VAI is clearly less mature but may evolve over time. The goal of VAI is to start off as a fully “governable” mechanism, launching with no rules and waiting for proposals by the community. 

Stability 

We have identified a few dimensions that can better put in perspective the spectrum of stability.

Frequency distribution of prices

We believe that possibly one of the most powerful diagnostic tools to characterize the stability of a protocol’s stablecoin is the probability distribution of its prices. In other words we would like a representation of the likelihood of the stablecoin reaching any price in the interval (0, +∞).

The probability distribution is however something we cannot precisely observe with a limited number of data points. We therefore obtain an approximation with the frequency distribution of the stablecoin price. To obtain it we divide the range of possible prices in a finite amount of intervals and compute how many times the stablecoin price fell into any interval. 

With the frequency distribution of prices, it is already possible to have a sense of the stability of the protocol. For example the number of observations within a certain distance from the one dollar peg can provide us with a simple measure of stability.

Furthermore, meaningful insights can be obtained from the shape of the distribution as well.

Notable examples are the distribution’s moments.

First and second moments (mean and variance) can give us an idea of the center of the distribution and its spread. Ideally we would want the distribution to be centered around 1$ and to have the lowest variance (and consequently standard deviation) possible.

The scaled versions of third and fourth moments (Skewness and Kurtosis) provide more subtle insights. Skewness can give us an idea of the distribution’s symmetry. If the prices observed are, for the most part, on one side of the 1$ peg, then there might be an imbalance in the design of the protocol. For example, if many more prices are observed significantly below 1$ than above it, we can infer that the protocol is not so effective in bringing the price up as it is in bringing it down. Kurtosis indicates the tailedness of the distribution; the higher the kurtosis is, the higher the number of observations in the distribution’s tails. A very high kurtosis might be symptomatic of the protocol’s fragility. In fact, it would indicate that when the protocol can’t maintain the peg the price starts to fluctuate wildly and reaches extreme values. 

Speed

Across the protocols in our research, we looked for models that react as quickly as possible to the external inputs that affect the dynamics of the protocols. We are not talking about governance response, which is always slow, but rather how fast the hard coded protocol model reacts to the market conditions and agent actions. 

Using the empirical data of the “Feedback Loop Index” introduced in token design, LUSD hard codes the feedback loop, resulting in the quickest response to stabilise the prices. 

Efficiency

Surely speed is important, but only if efficiency is not compromised. 

By efficiency we mean how efficient stabilization measures are. As there are always sacrifices that need to be made to rebalance the market, how well were the resources at the protocol’s disposal used to regain the peg. How accurately the mechanism in place helps reach the stability goal in terms of collateral usage and potential losses. 

The collateralization ratio comes into play here, as some protocols require a higher ratio than others to lower a systematic risk cascade. Users care both about safety of the derived stablecoin and the return on their invested funds so a balance needs to be met. 

There is something very important to consider in this regard: the underlying collateral. Backing a stablecoin with another stablecoin is totally different from backing it with one or multiple volatile tokens. 

Stability Mechanism: Direct vs Implied arbitrage protocols

We came across this dichotomy a few times during our research. Stability can be reached by continuous arbitrage opportunities and convergence to the peg (also called soft peg mechanisms). 

Protocols such as Syntetix have this feature as the stablecoin they created was just a byproduct and fits into the larger picture. The advantage of this is that there is an immediate use case and more demand, the problem is that other issues may arise before fixing an unpegged stablecoin and the solutions to problems may be exclusionary, especially with arbitrage. Syntetix handles this well and has been quite stable.

On the other hand, Liquity is a good example of a mechanism that lies on the other side of this spectrum. LUSD is indeed directly redeemable for Ether at face value when the price approaches $1. On the upside instead, the minimum collateral ratio of 110% creates a natural price ceiling at $1.10, since when the LUSD:USD rate exceeds that level, borrowers can make an instant profit by borrowing the maximum amount against their collateral and selling the LUSD on the market for more than $1.10.

Incentives

We have identified two major approaches to analyse incentives around collateralized stablecoins. One, of course, focuses on the recurring theme and fulcrum of this research: stability. The second one instead includes ‘mere’ incentives that do not contribute or help the stability of the protocol but rely simply on economic interests. 

Incentives and stability

Following the rationale mentioned above, we see that some projects (such as Synthetix and Alchemix) rely mainly on arbitrage while Liquity tries to intertwine economic incentives and stability by allowing users to gain rewards when contributing to the protocol’s stability pool. Indirectly though, also Synthetix incentivizes its users to deposit collateral contributing to the overall stability by compensating them with stakings rewards.

MakerDAO instead targets purely economic forces: demand and supply are theoretically influenced by the DSR and the stability fee. It is not easy to find a direct relationship between the effectiveness of those models.  

Pure economic attractiveness

Stepping back for a second, the whole DeFi is packed with examples of incentives that are not directly linked to the stability or the efficiency of the protocol (besides liquidity) but purely target user acquisitions and retention.

Liquity offers protocol fees to users who choose to stake LQTY. On Venus, depositors of collateral have been rewarded with the native token XVS for depositing collateral and minting VAI. In pure DeFi ethos, also staking VAI had historically a generous reward express in XVS.  

Overall, compared to other DeFi segments, collateralized stablecoins do not offer to users remarkable (and purely economic) incentives. 

Rewarding vs Punitive measures

We tried to analyse what type of measures can be more effective and if we had any evidence of punitive disincentives (FEI) being a viable solution to reach competitive stability. The market (and broadly speaking the human behavior in general) showed us that punitive measures can exacerbate negative feedback loops that bring agents to rush towards the exit door jeopardizing the overall stability of the protocol. 

Secondary Token

In all the protocols we have studied, a secondary token was present. Secondary tokens can cover a multitude of different use cases, some notable examples could be: governance, risk reallocation, incentivization and use cases linked to game theoretic mechanisms in the protocol.

Theoretically speaking, a secondary token could be avoided when designing a stablecoin protocol. Similarly to how Uniswap launched without a protocol token, a stablecoin protocol could function just with stablecoin itself. Furthermore, differently from what happens with most DeFi products, since stablecoin platforms necessarily entail the creation of a token, the protocol could achieve decentralized governance without the need for a governance token. The solution being to assign voting powers proportional to the amount of stablecoins a user holds. The stablecoin could become the de facto governance token. This situation where a protocol can use its main product as a governance tool is very much peculiar. 

If the secondary token is used for the Governance of the protocol, as it is the case in most of the cryptonetworks in considerations, one important question arises: what types of users should the protocol aim to target with the token? There exists a clear tradeoff between wanting to target the most active and involved individuals on the one hand and wanting to reach the largest number of people both for reasons of decentralization and aggregate demand.

When the secondary token is used to shift risks in the protocol, it is reasonable to expect that the party who takes on additional risks would be compensated fairly. Usually secondary tokens could be used to cover losses derived from risky collateral causing debt positions to go insolvent.  When designing a secondary token for a stablecoin system one should properly assess if risks are compensated with adequate rewards. In particular, it is extremely  important for the short term success and bootstrap of the protocol that the rewards offered are competitive in the yield markets of DeFi.  

If there are so many possible use cases for a token, one could wonder whether multi purpose tokens could be created. Designing a token system is certainly a complicated task, therefore a clear-cut answer is difficult to be found. The more functions get implemented in the same token, the more complex the system becomes as different mechanisms of the protocol could interact in unexpected ways through the token itself. For this reason combining different features in a single token is a delicate task and proper considerations on the complexity of the system should be made. On the other hand combining different use cases in a single token, when done in the right way, could simplify the system and make it more elegant, increasing the network effects and attracting a larger variety of agents.

Recommendation

Network Effects

LUSD scoring the highest overall when compared to the other stable coins could indicate that the exchanges for transactions are more popular and influential to users than the other stable coins. This also implies that they can disseminate information more quickly.

Although DAI has been one of the longest-existing stable coins that dominate the cryptocurrency market, one factor that may have been causing it to fare badly is that there could be too many means of exchange used to make transfers, causing most exchanges to be the least popular among users.

As a result, reducing or eliminating the least popular modes of transfer may aid in gaining influence over more well-known exchanges. Because LUSD, ALUSD, FEI, and VAI are newer than DAI and SUSD, it is possible that users will only be familiar with a subset of the popular protocols and will only be able to play with them for a limited time. As such, more recognition would be required for the newer stable coins in order for users to understand if these stable coins actually work well in the long run.

Bootstrapping Protocol

Stablecoins must not be an end in themselves. Of course, the liquidity required to create a market and give them the qualities of a currency must be there, but without a solid use-case to take them forward the projects have no after-hype future. Projects pride themselves with “partnerships” which are most often just DEX listings and don’t create an actual use case for the coin.This is why we suggest the following 2 stages of creating better network effects. Firstly, get a liquid enough market on a few of the major DEXes. 

Growth Stage of Protocol

Secondly, create partnerships that go beyond “exposure” and pure marketing. DEXes do not count in this stage. While it’s easy to look at sUSD and their great use case for a stablecoin, we understand not everyone has a whole ecosystem to build into. There are still plenty of projects that could use a stablecoin, your protocol just needs to show why it’s superior to the big fish and get transactions flowing. 

Demand

Stablecoins aim to bridge the divide between the digital asset world and fiat currencies by offering curbed volatility coupled with the benefits provided by the decentralized, secure, transparent quasi-anonymous payment system afforded by blockchain technology and cryptocurrencies. Protocols which have a more stable stablecoin price enable transactors to safely purchase and invest in other cryptocurrencies. In terms of stability, LUSD fared the best as compared to the other stablecoins, which suggests it makes a better base currency for trading on DeFi platforms.

The global financial economy requires a payment system where you can make payments without any delay, and that is cost-effective and free from intermediaries. Hence, it is important for stablecoins to minimise fluctuations in volatility to fulfil the need. 

Improving branding of stablecoins not only employ the fundamental aspects of price stability and blockchain technology used by today’s stablecoins, but also include loyalty program integration and secondary-market liquidity. Protocols could brand their token as a stablecoin for daily payment to stabilize the peg. Additionally, protocols could also work with each other to improve branding, which in turn promotes multi-party transactions and trading outside Dexes. Transactors could also benefit from partnerships through lower transaction costs, increasing customer loyalty. It is important for protocols to choose the right business partner as it’s not just the customer loyalty that matters, but also the knowledge and expertise gained. 

Marketing is a supplementary method to promote the stablecoins, along with branding. Through marketing, customers can get to know about the value of the stablecoins, usage information and additional info that might be helpful to the customers. Currently, the stablecoins are mainly interacting with their own protocol, and demand is volatile and constantly fluctuating. Branding and marketing will serve as platforms to increase the real demand, which in turn increases stability of the price peg as a whole. Only a real use case can bring about real demand that will last longer than the momentary bull run. As said before in network effects, such use cases can be brought about with partnerships that go beyond marketing and exposure, but actually integrate the stablecoin in a new DeFi protocol that needs such a service. We can expect many more such opportunities as more risk-averse investors enter the market which will value stability and predictability. 

Governance

Depending on the protocol, it is important to balance both hard and soft governance. The key reason is that voter’s apathy and fatigue are real issues. Majority of voters are unlikely to constantly vote for every single issue. If systems can be embedded dynamically as a hard coded governance for its resolution mechanisms, that can allow the community to focus more on improving the system in other ways than focus on solving the flaws. 

As governance is all about incentives, what we’ve seen is protocols going decentralized way too fast, when there is no faithfull community which has a significant enough stake in their success. The basic incentive for governance – a prosperous protocol, is thus not yet effective. Younger protocols are also much more likely to require more intervention as market conditions evolve and bugs are found. This is why we suggest teams not relinquish control until the protocol works very well and has survived at least one full market cycle. 

When the protocol matures we of course want to see decentralization, it’s the whole point of DeFi, but experience has shown that too many times DAOs are not effective at leading development. This is why we firstly suggest limiting the voted-upon issues to only big development decisions, excluding votes on small parameter changes. A compromise can be reached by having a representative democracy system when an elected council votes on procedural things (look SNX) and the big decisions are made with pure democracy. Such votes by the average users should be rewarded, either with secondary tokens or, preferably, the stablecoin itself. This gives an immediate use case and the users can better comprehend the value of their vote. The reward could be flat or relative, based on the number of tokens voting. 

There are two more scenarios worth talking about. As regulation develops, decentralized governance could become the only way to evade it. In this case, protocols will need to deploy a strategic plan of fast decentralization, especially if working from the US or Europe. If the community is built before a live protocol (which is difficult but possible) then all regulation problems could be avoided. Another option is joining an existing DAO structure (e.g., Raid Guild), but this does entail some cost to the protocol and existing users. 

The worst case scenario is decentralization and an exit scam. As developers usually cannot exit with stablecoins they might own, due to inherent price risks, selling secondary tokens too fast purely to make money in another, more valuable currency is a potential exit strategy for a malicious developer. As the level of due diligence is low and there are no regulations on selling such tokens, a sham protocol could be sold to clueless investors under the guise of decentralized governance. It is up to investors to judge if the governance effort makes sense and if the protocol is mature enough to survive the transition period. 

Stability

With respect to hard and soft pegs, we believe that no solution is necessarily the most adequate in any situation. Instead, different solutions might be appropriate in different scenarios. 

When the stablecoin is likely to be used mostly inside the protocol’s mechanisms, a hard peg system could be appropriate. For example in SNX, the token is used as a reference to obtain the price of synthetic assets in the protocol and is based on a hard peg mechanism.

When the stablecoin is used in the broad DeFi ecosystem, it is vital for it to maintain the peg effectively in market protocols external to its native ecosystem. In these situations a hard peg design is advisable.

With respect to efficiency, everything turns around a key parameter: the collateralization ratio. We believe that setting up an adequate collateralization ratio is a complex task, in fact it entails a tradeoff between capital efficiency and insolvency risks. Moreover the capital ratio is usually based on the risk profile of the collateral, represented both by its volatility and its correlation with the broader market. But these are not exact parameters as they require to be measured and measurements can be done in different ways. Furthermore both the parameters change over time, requiring adjustments over time.

For these reasons we believe that, in fixing adequate collateralization parameters, the proverbial wisdom of the crowds should be leveraged. In particular we believe that the parameters should be fixed either with a decentralized governance mechanism entailing proper incentives or, through some information generated by a market. On the mechanisms to leverage market efficiency some theoretical results might be relevant. Equations from control theory like PID feedback control or closed loop transfer function can be useful in aggregating the market data to update the system. 

A proper example of decentralized governance with right incentives can be found in the design of Maker DAO.

Incentives

Incentives are a broad topic in the design of a token system, with these recommendations we leave any claim to completeness. Nonetheless some general guidelines in the design of incentives were observed to be effective in fostering a healthy development of the ecosystem.

We identified three main ways in which users can contribute to the stablecoin’s stability. 

First, active users can take part in the protocol governance by making responsible decisions. Responsible Governance is already implicitly incentivized by the token value being tied to the protocol’s health. Furthermore, governance participants can be incentivized to act in the protocol interests with punitive measures. As an example, when a debt position goes insolvent in Maker DAO, the losses are covered by issuing new MKR tokens. This mechanism transfers losses on MKR holders, responsible for the collateral ratios, through dilution of the MKR token.

Secondly users can safeguard the stablecoin’s stability by holding the token even in distressed markets. If users are willing to hold the token in distressed markets, fluctuations in price do not get amplified by game theoretic considerations on the solvency of the protocol (bank runs). As an example we found many protocols, like FEI, to suspend the 1:1 convertibility of the stablecoin in a distressed market scenario. When the 1:1 convertibility of the token is suspended, tokens are redeemable at a fraction of their fair value, this disincentivizes users from dumping them and mitigates price fluctuations due to fear.

Lastly, users could contribute to the protocol stability by deciding to take over some insolvency risk. A system where users stake some tokens to cover insolvent debt positions could be designed. This way these users would effectively be selling insurance to the protocol and would be entitled to receive an adequate income generated by the protocol profits (often consisting in transaction fees).

Secondary Token

We believe that a secondary token is overall a better tool to have rather than not. That said, several best practices should be kept in mind during the design phase. The secondary token should:

  • be part of a broader picture stability and longevity is the main goal to achieve. That means that one important, if not the main raison d’être for the token would be its actual supportive contribution in the stability mechanism. That could be implied (indirect) through a staking mechanism for example or more explicitly as an insurance, last resort  buffer (à la Maker). 
  • have a balanced design of its incentive to meet an attractive early bootstrap plan and a long term, recurring incentive that warrants a certain, predictable loyalty from the holders. 
  • optimize for risk/reward so that users are fairly aware of the risks they would encounter. The goal would be to have those informed users managing their risks with available tools provided by the protocol itself. 
  • be designed to make the “protocol” aware of who their target audience is and be as reactive as possible to the potential behavioural changes of those agents. This can be done via Monte Carlo simulation or to provide more information, as discussed on the previous point 

Finally, on the governance function of the token, it is hard to give a recommendation without picking a side. Some of us believe minimal governance is actually a viable solution and it is working well in practice. Others believe that a natural evolution of the protocol is instrumental for the long term success of a protocol that needs to be seen as an integrative part of the overall, fast-moving DeFi ecosystem.

Conclusion

In this report we laid out the basics of how a collateralized stablecoin works, from minting to the users wallet and beyond. While there is still a long way to go before mass adoption, stablecoins already reflect many of the qualities exhibited by ordinary currencies. With the shift to digital fiat imminent, stablecoins should become an even more prominent decentralized alternative.

Most people would agree that stability is the most important feature of a stablecoin, it’s in the name afterall. We found that despite their reputation as very stable, collateralized stablecoins often deviate from their pegs by significant margins and maintain different shapes of stability with negative or positive skewness. It’s difficult to say which part of the overall currency model is most responsible for stability, most likely only a combination of all parts working in unison produces a good result. There is no question that DAI is the king of decentralized collateralized stablecoins with a precisely held peg, large daily volume and an active governance. The smaller experiments often lack one or more of these factors and thus cannot hold a peg in the long term, especially aftermarket shocks.

Secondary token was found to be an extremely useful tool in the design of a stablecoin for different reasons. In the first place a secondary token allows the allocation of governance powers to different users from stablecoin holders. Then a secondary token could be integrated cleverly in the system to provide different utilities. Some examples include: shifting risk from more risk averse to more risk loving users, bad governance punishment through dilution and good governance compensation through token rewards. Lastly the token can be used in the early stages of the protocol to bootstrap adoption through incentives programs.

The stabilization methods were almost all based on third party arbitrageurs rebalancing the peg. Many times the arbitrage relation is maintained with the collateral asset value through custom made conversion functions within the protocol. We found arbitrage strategies to be an extremely effective mechanism to maintain the peg of different stablecoins. 

While we may wish that all our code is perfect from the start, that is never the case. Governance and ways of altering and improving the protocol (in a decentralized way) is thus of vital importance. Without getting too deep in political science, DeFi faces similar issues as our societies with democracy not always working as intended. Incentives for decision making are not always sufficient or may even stimulate bad behavior, whales may control too much voting power and developers often release unready protocols into the wild to fend for themselves and develop without strong leadership or communities. Designing a really good and resilient decentralized governance system is very hard and it’s hard to say if it’s ever been done so far.

Appendix 1: Update

Liquity

In the months following the protocol’s study we did not find any major changes in the structure of the smart contract. Liquity does not have a system for decentralized governance therefore in order to change some features of the protocol the smart contract must be redeployed.

Synthetix ($sUSD)

Synthetix has been working on continual phased rollout of the Synthetix protocol on the Optimistic Ethereum Layer 2 solution. This solution will substantially decrease gas fees and reduce transaction costs for interacting with the protocol. During the month of July Synthetix Exchanges were deployed on Optimistic Ethereum L2.

With respect to Governance, the responsibilities and the role of the grantsDAO were formalized in the SIP-141. 

Fei

Fei Protocol launched $FEI, a stablecoin with $USD equivalent value, which is decentralised, liquid and scalable. This makes Defi platforms freely and securely accessible to everyone.

FIP-3 reinstates reweights on the FEI-ETH Uniswap pair with a fixed cadence. This removes the “incentive parity” trigger condition for reweights on the EthUniswapPCVController.

Included parameter changes:

  • Set reweight frequency to 4 hours
  • Lower the min distance below peg for a reweight from 1% to 0.5%
  • Lower reweight keeper reward to 200 FEI

FIP-11: Re-parameterize FeiRari: Update existing FeiRari Fuse Pool Parameters:

  1. Reduce Close Factor from 50% to 33%
  2. Reduce Liquidation Incentives from 10% to 8%
  3. Increase $FEI, $ETH and $DAI collateral ratio from 75% to 80%
  4. Update to FEI IRM
  • Reduce $FEI Base Rate from 2.73% to 0%
  • Reduce $FEI Initial Slope Multiplier from 36.40% to 7%
  • Reduce $FEI Second Slope Multiplier from 1650% to 400%
  1. Update to $TRIBE IRM
  • Reduce $TRIBE Base Rate from 2.73% to 0%
  • Reduce $TRIBE Initial Slope Multiplier from 53.75% to 20%
  • Reduce $TRIBE Second Slope Multiplier from 4000% to 400%
  1. Update to DAI IRM
  • Reduce $DAI Base Rate from 2.73% to 0%
  • Reduce $DAI Initial Slope Multiplier from 53.75% to 10%
  • Reduce $DAI Second Slope Multiplier from 4000% to 400%
  1. Update to $ETH IRM
  • Reduce $ETH Base Rate from 2.73% to 0%
  • Reduce $ETH Initial Slope Multiplier from 53.75% to 8%
  • Reduce $ETH Second Slope Multiplier from 4000% to 400%

Venus

The protocol has recently disclosed a comprehensive policy on risk assessment of the collateral accepted in the system: risk metrics include depth and volume level (which aims to measure the liquidity risk), volatility level, index level (measuring trust and counterparty risk), code commit level (which measures mostly counterparty risk) and contract holder level (helping checking the liquidity of the protocol). 

A second, but not less important update, was shared with the community in regards to a new interest rate design: the model is calibrated to manage liquidity risk and optimise utilisation. It follows the industry standards introducing a combination between linear rate and jump (kink) rate. The new model should sharply increase the solidity and robustness of the the protocol. 

Maker

In the last two months, two major milestones were reached.

In June the Liquidation 2.0 module was activated for all collateral types approved in the Maker protocol. This new module changes the auction mechanism used to liquidate a user’s collateral. From English auctions in version 1.2, the liquidation 2.0 module introduced Dutch Auctions which, unlike English ones, settle instantly.

On July 20 the full decentralization of MakerDAO has been announced. The Maker Foundation has fulfilled his bootstrapping responsibilities and is going to be dissolved in the following months.

Alchemix

Alchemix has no updates on the stablecoin protocol. They launched a new pegged token, $ETH.

Appendix 2: Intro to Protocols

In this report, we analysed 8 different types of algorithmic stable coins. 

  1. Liquity
  2. sUsd
  3. Fei
  4. Venus
  5. Maker
  6. Alchemix

Liquity

What is Liquity?

Liquity is a  decentralized borrowing protocol allowing users to  draw interest-free loans against Ether used as collateral. All loans are paid out in their native stablecoin LUSD and need to maintain a minimum collateral ratio of 110%.

Main Feature

Stability Pool and Liquidations:

The protocol introduces a stability pool which is filled with LUSD deposits from users: when any trove (vault’s name in the Liquity system) is liquidated, an amount of LUSD corresponding to the remaining debt of the trove is burned from the Stability Pool’s balance to repay its debt. In exchange, the entire collateral from the trove is transferred to the Stability Pool.

BaseRate: 

The protocol has no interest but introduces a dynamic fee (called redemption fees) based on the baseRate state variable in Liquity. This rate increases with each redemption, and decays according to time passed since the last fee event – i.e. the last redemption or issuance of LUSD.

Highlight

  • 0% interest rate 
  • Minimum collateral ratio of 110% 
  • Governance free (all operations are algorithmic and fully automated, and protocol parameters are set at time of contract deployment)
  • Directly redeemable at face value for the underlying collateral at any time
  • Fully decentralized 

sUSD

What is sUSD?

sUSD is the native stablecoin of the Synthetix protocol. It is minted through the protocol itself (through an app called Mintr) by staking the SNX native governance token with a C-ratio of 750% (changes depending on governance). It plays an essential role in the protocol, from denominating debt to collecting trading fees. It maintains its peg well by leveraging the Chainlink oracle. 

Main Feature

sUSD Stablecoin and Synths

There are 3 types of tokens in the SNX ecosystem:

  • SNX Native token — this token allows you to vote in the governance of the protocol, stake to accrue fees and participate in liquidity mining. It backs the issuance of sUSD.
  • sUSD stablecoin — can be used as an ordinary stablecoin and is traded on multiple DEXes. It is used to buy other synthetic products (Synths) such as tokenized equity, tokenized crypto and tokenized indexes. Exchange fees are collected and distributed in sUSD as well as rewards from staking SNX.
  • Synths — the main functionality of the SNX protocol, these can be tokenized assets of all types.

Highlight

  • Assumed to always be =1 by the protocol 
  • Decentralized governance by elected council
  • Fundamental role in a large and diverse DeFi protocol
  • Pure arbitrage maintains the peg

Fei

What is [Protocol]?

The Fei Protocol is a protocol whose goal is to maintain a liquid market in which ETH/FEI trades closely to the ETH/USD price. FEI achieves this via a new stability mechanism known as direct incentives. 

Main Feature

FEI is an algorithmically stablecoin pegged to $1. That means $FEI’s economics is determined mathematically to get the $1 peg. 

Direct incentive stablecoins use dynamic mint rewards and burn penalties on DEX trade volume to maintain the peg. FEI uses Uniswap as its incentivized DEX at launch. Governance can add and update DEX integrations and other incentives as needed.

$TRIBE is a governance token to adjust the parameters to change the mechanisms of $FEI.

The protocol aims to reach two main objectives:

Objective 1: FEI proposes a new stability mechanism called direct incentives. A direct incentive stablecoin is one in which both the trading activity and usage of the stablecoin are incentivized,where rewards and penalties drive the price towards the peg.

Objective 2: the key innovation behind the Fei Protocol’s mechanism is the idea of Protocol Controlled Value (PCV), a subset of Total Value Locked (TVL) in which the protocol outright owns the assets.

Highlight

  • Decentralized governance with the possibility to delegate votes
  • $TRIBE Governance token
  • Direct Incentives Stability Mechanism
  • Dynamic mint rewards and burn penalties depending on the current peg

Venus

What is Venus?

Venus is a money market protocol launched and supported by the Binance (Smart Chain) ecosystem. The vision behind its creation was to initiate a protocol fully owned by the community and for the community. It started on the Compound + MakerDAO template with the idea of iterating via EIPs on those basic guidelines. 

Main Feature

Venus enables users to utilize their cryptocurrencies by supplying collateral to the network that may be borrowed by pledging over-collateralized cryptocurrencies. The lender receives a compounded interest rate annually (APY) paid per block, while the borrower pays interest on the cryptocurrency borrowed. Interest rates are set by the protocol in a curve yield. Venus allows to use the collateral supplied to the market

not only to borrow other assets but also to mint synthetic stablecoins with over-collateralized positions that protect the protocol.  

Highlight

  • Launched with little to no rules and features and added parameters via governance votes. 
  • Fully decentralized governance  with no backers and pre-sales. Funded by Binance Smart Chain ecosystem and managed by Swipe. 

Maker

What is [Protocol]?

MakerDao is a decentralised lending platform that takes crypto collateral to issue a crypto-backed stablecoin called $DAI, it allows different collateral types and  is governed by Maker ($MKR) holders.

Main Feature

The system allows its users to deposit different types of collateral to obtain loans denominated in $DAI. Leans are overcollateralized with different ratios depending on the type of asset posted as collateral.  Once a debt position goes undercollateralized (< collateralization ratio %) it is liquidated by third party users and the borrower is penalized.

When a debt position goes insolvent (<100%) effective losses are generated, the protocol covers them by auctioning newly minted MKR tokens. 

The MKR token entitles its holders to receive a stability fee of about 1%. Furthermore it is used in governance to vote on different proposals. As an example of governance the $MKR token is used to set collateralization ratios for different collateral assets. 

Highlights

  • Multi collateral debt mechanism
  • Different capitalization ratios set by decentralized governance
  • 1% stability fee
  • $MKR governance token
  • Insolvency resolved with new $MKR issuance 
  • Collateral redeemable only by the original borrower in normal times
  • Collateral can be liquidated by third parties when the position goes undercollateralized

Alchemix

Alchemix is a platform for the creation of yield-backed synthetic tokens, in our analysis will focus on the stablecoins,  that users can acquire for no cost in exchange for locking collateral in the system. 

The stablecoin synthetic token (alUSD) is basically backed by future yield earned by the collateral deposited in the platform. 

Main Feature

User needs to deposit DAI with a 200% collateralization ratio requirement into a vault. Users’ DAI is then deployed  to earn yield in the yield strategies platforms. Users can mint the stable coin alUSD up to 50% of the amount of DAI they have deposited. The  yield generated by the strategies is then used to balance the existing asUSD debt, which will slowly decrease with time. As the protocol repays users’ debt, they can withdraw more DAI or mint more alUSD as desired. Once a user has 0 alUSD debt remaining, they are free to withdraw all of their deposited collateral. 

Highlight

  • The Alchemix protocol. In other words, the loan pays itself off.
  • 200% collateralization

If you liked this report, consider following us on social media: