Maker's problems stem from a chronic lack of collateral throughout the system, the dark Web silk Road founder wrote in prison

By Ross Ulbricht.

Original title: "Analysis and Improvement of DeFi Protocol released from Prison by Founder of Silk Road"

Translation: Babylon

I recently read the Maker Agreement white paper. What a cool concept! & have spent The people behind Maker have created a cryptocurrency that tracks the value of the dollar (stable coin). You'll get all the benefits of cryptocurrency without wild price swings. Naturally, it has become very popular and has created hundreds of millions of dollars in value.

I took a closer look at Maker's documentation and read some articles about its recent collateral crisis in March 2020. I can't fully understand the system and what's going on inside the prison, but I can see it on the surface and I believe I can come up with ideas to help prevent this kind of crisis from happening again.

Maker's agreement issues stable COINS called DAI.

& have spent

DAI 

DAI's value is backed by collateral in a virtual vault on the Ethereum blockchain. The owner of the safe deposit box is the foundation of the whole system. The value in their vaults supports the value of all DAI in circulation. If the total value in the depository has ever been below the dollar peg value of all DAI, the system is bankrupt and can be considered a failure.

I think the problem with the Maker agreement that led to the recent crisis is a misunderstanding of the role of the vault owner. In several places, I called them "borrowers," which didn't make sense to me at first. What are they borrowing? Then it sounded: the idea that the vault owner was borrowing DAI from the Maker Protocol itself and placing collateral (usually in ether) in its vault to support the loan. It's like taking out a mortgage on your house, ether is the house and DAI is the bank's loan. Like collateral, vault owners must also pay interest on borrowed money, a so-called "stability fee". And if the value of the collateral falls too low, their vaults can be auctioned off.

It's like a house going into a foreclosure. That is why the analogy does not work: DAI "borrowing" has no value other than the collateral backing it. Its value comes from the fact that, when pushed to the top, it can redeem its collateral in a liquidation auction. Maker Agreements already act as Banks, issuing mortgages, but to me, the vault itself is the bank, or at least the bank used to be.

Before the advent of modern central Banks, Banks kept gold in vaults, then printed and lent paper money backed by the gold. & have spent The owner of the safe deposit box in the Maker Agreement is doing something similar. They keep the ether in their vaults, and then lend DAI, the currency's backing.

As a result, they are more properly treated as lenders than as debits. Banks should keep enough gold in their vaults to cover all currencies in circulation backed by gold. In practice, they print more money than they can afford, just hoping that everyone doesn't immediately show up with their gold. When the inevitable bank run did occur, bankers turned to their friends in government. Now that the dollar has been wiped out, sporadic reserve Banks have become the norm. The Banks are still in trouble and could be bailed out. It's a little messy.

This is not a good model for Maker.

The agreement is not a modern bank because it has no government backing. Instead, the custodian owners should be treated as old-fashioned, old-fashioned Banks, but they cannot cheat because all the accounting is done publicly on the blockchain. If they try to issue MORE DAI than their Ethereum can support, they risk liquidation. The fundamental difference between this and the design of Maker's agreement is that, as a lender, the owner of the vault should receive interest, not pay it.

When I read the paper, my first question is "why do you want to charge to others, why someone put their precious etheric COINS in a vault, to support the stability of the coin (DAI) others? In addition to the stable fees, if the value of the mortgaged property drops too low, they may also lose its collateral 13% liquidation costs. So why take the risk to assume all the risk?" It turns out that the system rewards other behaviors (it should not) and that the repository owner can take advantage of it.

The agreement has smart contracts where users can lock in their DAI and get a "savings rate". & have spent 'All right,' I thought. & have spent "If the savings rate is higher than the stability fee, then the owner of the vault can cast DAI backed by ethereal in the vault, put it in a savings contract, and - as long as they pay attention to the liquidation rate - receive a return in ethereal (savings rate minus the stability fee)."

But why should we motivate this?

The whole point of DAI is that it can circulate as dispersed, blockchain-based dollar tokens rather than being locked into smart contracts.

It turns out that before the crisis in March, the savings rate was about 8 per cent and the stabilisation fee was about 0.5 per cent. My guess, therefore, is correct: rather than selling DAI and becoming a good steward of their vault, the owner of the vault will be paid (7.5 per cent), but will keep the DAI he has forged and deposit it in a savings contract. Clearly, before the crisis, a significant portion of all DAI that existed were savings contracts. (See Figure 1.)

The purpose of the savings contract is to enable the committee overseeing the agreement (the manufacturer DAO) to adjust savings rates so that DAI flows in and out of circulation, affecting prices and keeping them at parity with the DOLLAR. I really don't think it's necessary. Market forces are what keep prices stable, as both sides of the parity have arbitrage opportunities to bring prices back to equilibrium. The savings contract is like a tool the central bank wants, so he can imagine himself "managing the economy". This top-down impulse to control prices does not sit well with the decentralised spirit of the agreement. Worse, incentive behaviour reduces the system, not stabilises it.

Fully decentralized protocols are "managed" by mathematics and logic, giving human decision-making power to end users. Relying on the Central Committee to set important parameters within the system is a weakness, not a strength. This is not to say that the Maker agreement has to work without a governing council. It's a complex system, and it's really impressive how it works. However, the essence of the Commission must be understood: expediency. It is not a valid mechanism for discovering the best values for the system parameters it monitors.

Crisis & have spent

Around mid-March, the price of Ether suddenly dropped by about 50 per cent, while DAI soared by more than 10 per cent (a large sum for a dollar one-to-one trade). As the value of its Ethereum fell and it was forced into liquidation, many vaults became collateralised. This is the risk that the owner of the vault is taking. Their job is to make sure that the ether is added to make up the shortfall or to pay off their outstanding DAI balance. But the crisis has become so severe that some vaults have been forced to liquidate at steep discounts, losing almost all, and in some cases all, collateral.

The problem stems from the fact that the system is chronically short of collateral. Not only are the vault owners penalized by the stabilization fee for using the system mortgage, but the savings rate is increasing DAI's burden. Then, during the crisis, no one was willing to give up its stable, high-yielding DAI in a liquidation auction in order to take the collapsed, negative yield ether.

Finally, DAI must be backed by the USDC, a stabilisation currency backed by bank cash (hardly a decentralised solution). If the Maker agreement is to be truly decentralized and independent, it must acquire basic incentives and resist the temptation of top-down control.

& have spent

A cautious suggestion & NBSP;

Clearly, I am in favour of abolishing savings contracts and stabilisation fees, and replacing them with what? First, there is no need to change savings contracts. & have spent DAI is valued for its stability fees, not arbitrary rates of return. As for the stabilization fee, it should be negative. That is, the owner of the vault should be rewarded, not punished, for providing collateral. Therefore, it should not be called a stabilization fee at all. Indeed, it should be called the savings rate, since vault owners keep their ethereal COINS in their vaults and receive returns to support DAI. But this can be confusing because the term is already in use. Instead, call it the collateral rate, the rate of return on collateral.

Whatever its name, it will attract collateral into the system rather than reject it. Where, you may ask, will DAI pay the mortgage rate? It should be issued by DAI from the vault providing the mortgage. DAI holders are offered services (decentralised stabilisation currency), so they should pay. And how to determine the ratio? A simple solution to the current system of MakerDAO (where the stability fee and savings rate have been set) is to reduce the savings rate to zero and the stability fee to a small negative number, perhaps -1 percent. This must be done so that DAI disappears at just 1 per cent a year, leaving behind far more collateral than would otherwise be the case.

A better solution would be a system in which vault owners could set their own rates and compete with each other to win DAI holders and win with the lowest rates. This usually keeps interest rates low, but when there is a severe shortage of collateral, interest rates automatically rise through market forces, encouraging more collateral to enter and determing DAI hoarding. This is the opposite of what we saw in the crisis of mid-March.

With appropriate adjustments in incentives, DAI should have a deeper market and tighter bid-ask spreads. The reason is that, at a positive rate of return, vault owners will want to put their newly created DAI on the market, so that others will hold depreciated assets (interest rates will be abolished when they own them) & NBSP; DAI). So there will be plenty of liquidity above parity.

Below par, as long as DAI holders can redeem DAI collateral, there is an arbitrage opportunity. If the price is below par, people can buy the discounted DAI, redeem it from the issuing vault for extra collateral at Oracle's rate and sell the collateral in full on the market. Oracle rates are the way the Maker agreement determines the value of collateral. So there will be plenty of liquidity below parity.

These forces will stabilize prices and ensure full collateral, especially in volatile market conditions.

& have spent

The technical part & NBSP;

The mechanism by which market-based systems determine mortgage rates is somewhat complex. If you're not interested in the details, skip to the conclusion section.

As mentioned above, vaults should set their own rates, and low-rate vaults can win the competition. But unless there is too much collateral in the system, there is no competition. If all collateral is backed by DAI, the depository owner can charge whatever fee is required, and redemption is the only resource for DAI holders. We can solve this problem by allowing unencumbrance to collect interest on DAI from another vault if the interest rate on its collateral is lower than that from another vault. So instead of casting DAI and selling it on the market, vault owners can collect interest at a lower rate. This is useful for keeping interest rates low, but there is no incentive to actually issue and sell DAI now, especially since there is a real risk of clearing the vaults.

A compromise must therefore be reached, whereby excessive collateral in the low-interest vaults captures only a fraction of the interest on the issue vaults and the rest is paid at the issuer's vaults but at the rate of the low-interest vaults. Therefore, in the case of excessive collateral in the system, there will be a natural write-off rate, above which the depository of DAI issued will lose the depository of excess collateral whose collateral is lower than the credit, and both yield lower.

Which vault from which vault? A vault with a ceiling of the lowest rate shall be captured by a vault with a ceiling of the highest rate. This will incentivize high-rate repositories to keep their rates low.

What part of the interest on the issue vault should be captured by excess collateral below the threshold? This should vary according to how much excess collateral is in the system and the total amount of DAI outstanding. If all DAI pay interest on excess collateral (that is, only the vault for excess collateral is below the threshold), the capture rate should be 0 per cent. The system is riddled with excessive collateral. The ratio should increase linearly, reaching 100 per cent when there is no excess collateral below the threshold in the vault: the system is close to bankruptcy and needs to provide excess collateral quickly.

All of this solves the incentive to keep excess collateral and low collateral rates in the system, but the reason we want excess capacity is so we can use it when we need it. To do so, a mechanism is needed to switch DAI automatically from a near-bankrupt vault to one with too much collateral. This can be done in the usual way by clearing auctions. The safe will be cast against DAI with surplus collateral and used for bidding. But it would be better to automate this to speed up the crisis and simply move assets and liabilities (collateral and DAI) into well-capitalized vaults. The depository may mark and charge a portion of the excess collateral for this purpose, and the depository shall first be liquidated as a low-cost depository.

The collateral rates set by the vaults in the system tend to converge to a simple "market rate", or cut-off rate. Repositories below the dividing line are "keep the money on the table" because as long as they don't cross the dividing line, they can charge more without losing out to cheaper repositories. On the other hand, vaults above their critical value lose too much collateral to vaults below their critical value.

When the collateral (or the value of the collateral in the event of a change in the exchange rate) falls, the cutoff value automatically rises, because there will be less excess collateral below the cutoff value, allowing the Treasury with the higher interest rate to charge the full amount. Capture rates will also rise because a small group of DAI is paying interest on excess collateral. Both would ease the shortage by encouraging new libraries and new collateral to flow into the system, all without MakerDAO's intervention.

I tried to consider a way that the capture rate could also be dispersed and controlled by market forces rather than by the above linear formula, but it could not. The problem is that the vaults are already competing for DAI holders' benefit based on the collateral rate, so how can they compete for DAI holders' benefit based on the capture rate? One could devise a two-parameter formula weighing these two values to make a case for interest in the vault with the lowest score, but this seems complicated and fraught with loopholes. One can also leave the collateral interest rate and capture rate to the Treasury, and leave the interest directly to the Treasury with the lowest total income. However, the Treasury can use extreme settlement methods to favor DAI holders with very low collateral rate, but can capture 100% of the collateral.

& have spent

Conclusion & have spent

The Maker Agreement is a very cool concept and I hope it succeeds, but I fear that if these fundamental issues are not addressed, it will again face the kind of crisis we had in mid-March 2020. The next crash could lead to systemic bankruptcy and catastrophic failure. It involves a lot of money, and even more potential. The crisis should be a wake-up call to reform or try a New Deal. I just hope these ideas will help point in the right direction.

Text: Https://medium.com/@rossulbricht/remaking-the-mak9f11

Translation Address:https://www.8btc.com/media/614578

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia