James Talks | What is decentralized collateralization?

Anything, as long as it has some relationship with decentralization, will immediately become ambiguous. The same is true for common mortgage lending and mortgage operations. The traditional mortgage is fairly simple. You give me a valuable asset, and I lend you some cash or other assets. When it expires, you will repay the principal and interest. I will return your collateral. If you don’t repay it, I will have the right to own or auction off your collateral. If the price of the collateral fluctuates dynamically, I have the right to ask you to add more collateral during the period that the price changes. Otherwise, I don’t have to wait until it expires before I can dispose of your collateral. This process is simple, clear, and everyone accepts it.

The following is our experience. The difficulty of matching and the time cost of both borrowers and lenders make p2p matching very inefficient. This is also the reason why most of the DeFi lending protocols in the early days failed. So everyone began to look for a way out. One solution is to build a pool with high capital utilization, and the other is to directly lend the borrower stable coins with their collateral.

The former is also known as the Compound model, and the latter is known as the MakerDAO model. Both models guarantee either lenders or borrowers exist in the system in the forehand, then seek the counterparty of the transaction.

Compound ensures that the lender’s capital pool is sufficient, while MakerDAO prioritizes the borrower’s asset pool. One is investment-oriented, the other is loan orientation. But if the whole process is on-chain and proceeds in a decentralized manner. Then the process above immediately raises the following concerns: First, who will complete the matching of borrowing and lending first? Second, once the price changes, who will ensure adding margin or liquidation?

We first built a typical matching lending model: the p2p model, where the borrower’s and lender’s needs were written in the contract. On this matching platform, once the transaction is concluded, the lender has the right to conditionally liquidate the collateralized assets.

This model looks reasonable, and it is completely decentralized, but new problems also appear. There are too many variables such as the borrowing term, interest rate, and collateral assets, etc. Matching is extremely hard. In other words, if the opportunity cost or reinvestment cost is too high, the actual rate of return will be small, and the willingness to lend will be gone. On the other hand, if the borrower cannot find matching funds for a long time, it will also withdraw from the market. , Causing the entire scale to shrink further.

In the two fund or asset pool models, the liquidation risk caused by the fluctuation of the mortgage price is always most important. If the liquidity of the collateral assets is infinite and the price is always valid. Then regardless of the collateral ratio, as long as the value of the collateral is higher than the loan amount, there will never be any liquidation risk.

Whether the collateral ratio is 60% (the scale of interest-bearing borrowing divided by collateral value) or 90% will always be the same. However, liquidity will neither be infinite nor will the price always be effective. In some extreme cases, price fluctuations directly cause the value of collateral assets to be lower than the value being borrowed.

Because everything is decentralized, you can’t lock the borrower to request it to pay for the loss. Therefore the lender will be the only one suffering, which will increase the risk of a bank run. In addition, in the fund pool model, liquidation cannot be completed by a designated person, because this will lead to a certain trust risk: the delay or pause of liquidation will cause greater losses.

Therefore, the liquidation process must be decentralized, which means that the design of the incentive mechanism must ensure that any third party has the motivation to do this. Among the current DeFi projects, the liquidation residual value is designed for this incentive.

Liquidation risk is typical tail risk. In the design of Compound and MakerDAO, this problem is solved by making everyone bearing the risk together. This will lead to an increase in the risk of the lender and the instability of the intrinsic value of DAI (no longer worth 1 USD).

For this kind of tail risk, the best solution is to introduce a decentralized insurance mechanism: in most cases, insurance premiums are charged, and compensation is made when such extreme situations happen. The entire process is decentralized. Anyone can become a participant of the Insurance Fund, and can also withdraw what they put in from the Insurance Fund Pool.

The role of the Insurance Fund is equivalent to the mortgage operator selling the tail risk to a third party, which can ensure the safety of the lender or the stability of the DAI. This design represents part of the borrowing interest and the so-called stability fee, which should be paid to the Insurance Fund. Consider that the scale of Insurance Funds is dynamic, so we can determine the relative value at market risk through the dynamic insurance scale! This provides an indicator to look at the liquidation risk of the entire pool.

As mentioned earlier, if no one is borrowing from the pool, it will cause the lender’s overall return to drop. This is what we call reinvestment risk: after receiving interest for a while, the lender will need a new opportunity to receive interest, and this is not considered as a long-term and stable return.

Combining the liquidation risk and Insurance Fund we just discussed, in this model, the Insurance Fund loses continuous stability fees or interest because part of the collateral assets is liquidated. This implies a term structure with different liquidation times: the earlier the liquidation, The greater the reinvestment risk, and the later the liquidation, the lower the reinvestment risk.

Under a given price model, the duration of touching the liquidation line is related to the collateral ratio or the liquidation line. Considering that the two often maintain a linear relationship, we use the collateral ratio to measure the time composition of the duration of touching the liquidation line: the higher the collateral ratio (borrowed assets divided by the collateral assets), the shorter the duration of touching the liquidation line. The lower the mortgage rate, the longer the duration of touching the liquidation line.

In this way, the collateral ratio creates a measure of reinvestment risk, which means that the insurance rate can be set based on the collateral ratio. For example, a simple solution is a linear formula. Of course, we can also apply a more precise formula if we want to be more accurate.

In this way, the Insurance Fund, and the insurance rate related to the collateral ratio (the insurance rate in Compound is part of the interest rate, and in MakerDAO, it is the Stability fee), creates the most basic risk management plan for decentralized collateralization lending or collateralization stable coin. This design is basic but atomic. We will see more and more lending projects to make such adjustments in the future.

Now let’s talk about revenue. Although the risk of lending out USDT by collateralization and generating DAI by collateralization are similar, the returns are worlds apart. The basic need for a stable asset is nothing more than investment and consumption (utility).

Let’s talk about consumption. USDT is more accepted on-chain. Whether it’s financing, on-chain quotation, or pricing on DEX, USDT has the upper hand. While off-chain, centralized exchanges set USDT as the denominated currency.

On the other hand, DAI is very unlikely to get the same treatment. There are many reasons for this, including the consensus, the difficulty of scalability, etc. In addition, if you enter the world of fiat currency, USDT has a currency reserve of 1:1 (approximately) with the US dollar for redemption, while DAI must be traded into USDT or converted into US dollars by some OTC service providers. The cost of this conversion and the scale of liquidity is far inferior to USDT. Therefore, from a purely use perspective, the increase of demand for DAI generated by collateralization heavily relies on the cooperation of various on-chain protocols, which is what everyone calls the “DeFi Lego”. This is more of the second demand we mentioned: investment.

Since DAI itself is a kind of derivative, and there is no particularity. If we copy a contract to generate a DAI2, it is almost the same as DAI.

From the perspective of the game theory, if the collateral price is 100 at the beginning, the Max collateral ratio is 50%, and 50 DAIs are generated. When the collateral price drops to 80, the second person continues to do so and he can get 40 DAIs. At this time, each collateral supports 45 DAIs.

But if the second person copies a DAI contract and generates DAI2 according to the same rules, at this time, each collateral supports 40 DAI2. So for the third person, the same asset, is he willing to pay 1 dollar to get a DAI which is backed by collateral that supports 45 DAIs, or get a DAI2 which is backed by collateral that supports 40 DAI2s? Without other additional variables, this choice is undoubtedly DAI2. So if DAI wants to attract users and let them ignore the scale of backed collateral, what needs to be done?

The first thing is to increase the OTC service to ensure that DAI can be converted into USD at any time. This path treats DAI as an intermediate voucher. What the Maker community does is provide USD loans or USD liquidity.

The second thing is to find ways to increase the application of DAI, such as using it in yield farming to gain excess revenue. Obviously, why do others allow you to obtain this part of the value? Why do people use DAI when they can use USDT? These problems completely depend on the value provided by the Maker team in the development of other DeFi contracts.

But this value is not stable, because the contract can replace DAI at any time, or even write a DAI2 by yourself. This harmonious picture is actually very fragile: one system always needs someone to provide value, and if one doesn’t do anything, but only goes to other systems to grab their benefits, it cannot last long.

Therefore, there must be some basic and solid demand to support the intrinsic value of collateralized stablecoins. From the perspective of the Maker community alone, we can’t see this kind of demand. As mentioned earlier, because it is too easy to be copied, lacks scarcity, and self-enhancement properties, it cannot constitute a unique on-chain asset.

Secondly, there is no starting point of this kind of demand within the community: What project needs dai, and does it have to be dai? This kind of question cannot be answered inside the Maker community. DAI is not like USDT, if the scale of USDT is expanded and then copied, the credit level cannot be guaranteed to be the same, while the DAI is completely the same because it is trustless.

A more reasonable idea is that the collateral used to generate DAI itself must complete a certain logical closed loop and DAI will be essncial to the loop. This constitutes a basic supply-demand relationship: collateral-DAI-the generation of collateral. In such a scenario-based collateralized stablecoin logic, we only see that Parasset has this consideration: the pUSD and pETH by collateralizing NEST are effectively used in the quotation system to generate more NEST tokens or NEST pricing quotations.

On the other hand, the use case of DAI has no such correlation and dependence, and cannot systematically reduce the capital pressure and asset fluctuation risk of NEST ecological miners. Moreover, the pUSD backed by the most NEST will be used first, which involves the difficulty of manipulating and affecting the price, which indirectly forms the self-enhancement attribute. Maker that is completely isolated from ETH is just a simple standard contract (the copy will be the same as the original).

If the demand is one by one concentric circles, the mortgage operation needs to construct the bottom concentric circles. Especially the collateralized stable coin, which must be derived from the inherent demand of the collateral itself to differentiate from various “DAI1”, “DAI2” and form self-enhancing properties.

This stable demand provides a clear insurance rate income source for the Insurance Fund, thereby forming the starting point of demand and liquidity, without the need for the project to perform OTC services (this is essentially centralized).

When this bottom concentric circle is formed, it will gradually spread to the use of some projects within the ecology, such as trading, lending, and derivatives, and then gradually spread to cooperative communities, general communities, and off-chain markets. The basis of these basic logics determines the long-term rationality of a blockchain project and continues to gain beneficial information during the long-term evolution. What success is to compare how much information WILL be beneficial to it in the FUTURE, not how much noise it gets TODAY. This is what we believe.




Parassets such as PUSD, PETH, PBTC are generated by protocol collateral and anchor the underlying asset with an intrinsic value of 1:1 http://t.me/parasset_chat

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

Why T&T crypto investors need a pull-out game…

Joseph Lubin

Advanced Crypto Matching Engine Solutions for your Cryptocurrency Exchange platform

SkyRocket is the official broker of Binance! 🥳

What I learned from Bitcoin.


Stable coin Marketcap tops $50 billion

10 Things Most People Don’t Know About bitcoin

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store


Parassets such as PUSD, PETH, PBTC are generated by protocol collateral and anchor the underlying asset with an intrinsic value of 1:1 http://t.me/parasset_chat

More from Medium

Uncles can be a wonderful thing!

Uber Chronicles #33

How prevalent is violence in Tanzanian schools?

Bias in Algorithms