Lin Rong: Three major problems and advantages of decentralized finance (DeFi) compared to traditional finance

lead

textThe Transformative Significance of DeFi）。

However, it cannot be ignored that many incidents have occurred in DeFi this year, from flash loans being used to attack bZx, to MakerDAO’s poor liquidation that caused many ETHs to be auctioned at 0 yuan, and the digital currency market was bloodbathed on March 12. Compound once had to rely on VC injections continued, leading to a $25 million theft from dForce in April.

Undoubtedly, this field is still very young and there are many gaps and shortcomings. It claims to revolutionize traditional finance, but it lacks sufficient attention to risks, respects tradition, and still has a big gap with the regular financial army in terms of data sources, price verification & emergency handling, real-time liquidation, and liquidation processing. ;

The choice between centralization and decentralization and the design philosophy of DeFi projects still need to be refined. Of course, DeFi still has the possibility of reform and innovation for traditional finance. The latter needs to learn from the former, just like the situation when FinTech was born.

In a word, DeFi needs to lower its so-called "reformer" attitude, and concentrate on learning mature and stable paths, methods and models from traditional finance, while traditional finance should be further open and compete with DeFi in a broader sense .

The third phase of Omni-Zhongcai Digital Economy Forum specially invited [former vice president of Deutsche Bank’s global marketing department and former senior expert of Ant Financial] Ms. Lin Rong to discuss the topic of bridging DeFi and traditional finance. The following is the transcript of the lecture, Enjoy !

secondary title

At present, most of the employees working for each of these "banks" only have a few to twenty or thirty employees, but the "savings balance" has grown rapidly, reaching a maximum of US$1 billion in mid-March. It quickly recovered to about 700 million US dollars. The number of users has also grown from zero to 50,000 to 60,000.

That may not be a huge number for traditional finance, but it needs to be taken into account that blockchain finance is still in its very early stages - only a little over 2,000 of the world's population of over 7 billion are estimated to own cryptocurrencies Ten thousand. The growth trend of the number of users is close to that of the Internet in the early 1990s.

Let’s take a look at the dynamics of four representative projects in the first four months of 2020, and discuss the advantages and unresolved problems of decentralized finance compared with traditional centralized finance. These four projects are all involved in lending business that we are familiar with, and they all deploy contracts on the Ethereum (ETH) blockchain network.

January: 15-second loan: Aave Flash Loan launched

In traditional finance, when a bank grants you a loan, it usually either investigates your situation and confirms that your credit score passes, or you have to provide collateral.

In flash loans, neither of these two items is required.

Even if you are a bankrupt person, you can lend a large amount of funds to a "bank" such as Aave, and then transfer and use it at will. How does the bank guarantee your repayment? You need to repay the money before the next account entry in the bank network system, otherwise the action of borrowing money and the things you do with the money will be declared invalid by the system—yes! Anyway, the money and your operations are all on the Internet, which is the system registration.

The "banks" we mentioned all use the Ethereum network system, which registers a block in about 15 seconds to record a series of transactions. This is a loan in about 15 seconds.

"Flash Loan" has been running like this for a while, and it became famous all of a sudden in February because of a clever attack.

Someone found a loophole in a leveraged trading protocol called bZx. The step of checking the latest price before placing an order to see if your margin is enough will be skipped: that is to say, you can use less money to increase your assets price, so the attack began. For a specific analysis, you can read this article "PeckShield: Hardcore Technology Analysis, The History of bZx Protocol Attacked by Hacker Vulnerabilities".

To put it simply, he first borrowed a large amount of money (that is, ETH, which can be said to be the standard currency on the Ethereum network) with another project that provides flash loans, and part of it was pledged for lending and hoarding a small-scale asset WBTC, a part of the above-mentioned problematic platform is leveraged to buy WBTC to increase its market price. Then, he took the opportunity to sell the WBTC he had hoarded at a high price, and returned the flash loan, leaving a profit. Such a simple operation is actually completed within 15 seconds, and within 15 seconds, this person has made a profit of about 350,000 to 60,000 US dollars.

February: Pool Together gets investment

The PoolTogether project thought of such a game: Let you buy a lottery ticket for one dollar, and promise to keep the capital: After a few days, the original money will be returned to you. What kind of prize can such a lottery win? The total amount of all lottery tickets sold will be deposited in the bank to earn interest. When the draw day arrives, all the interest will be taken away by the winner at random.

Can I have a digital "dollar" to place an order online? Yes, they have integrated USD stable currency MakerDAO’s Dai and USDC, which are 1:1 pegged to the real USD.

What is the concept?

For example, when you are alone and no one cares about you, you can access JPMorgan Chase Bank's electronic check system, Citibank's deposit service system, Bank of China's exchange service, and the audit services of the Big Four accounting firms... They provide Your project provides support, telling your customers the ins and outs of every transaction, open, transparent, and traceable.

The basic service modules that are almost monopolized and closed in the traditional world are now deployed on the blockchain, like building blocks, open for your use, so you can use the smallest resources to unleash your creativity, combine various ready-made modules, and build a Various forms.

March: Buy ETH for 0 yuan, MakerDAO lost $4 million

The MakerDAO project is a classic "mortgage loan" logic. For example, if you mortgage a house, you can borrow cash at a discount on the house price, or a more similar traditional financial tool is repurchase, which mortgages stocks, treasury bonds, gold or copper and other commodities to borrow money.

To put it simply, in MakerDAO, you can pledge digital currency such as Ethereum ETH, discount the market price, and lend out the US dollar stable currency Dai issued by the project itself. The price target of each Dai is equal to one dollar.

In order to prevent the price of the pledge from falling too much and not enough to repay the loan, the rules set that if the market price of ETH falls to a certain level, you will lose the pledge if you fail to add more pledges or repay the loan in time, and the system will let others take it away and auction it , to repay the loan.

On Black Thursday, March 12, the price of ETH plummeted, and some orders entered the collateral auction process. As a result, about 4 million US dollars of ETH was auctioned off at 0 yuan.

What went wrong? We only need to look carefully at the description of the rules sentence by sentence, and we will know that in addition to the obvious interest rate and discount rate of this type of product, there are three points that need special attention in terms of safety.

The first is the "market price" - how the market price is determined, or in blockchain it is called an "oracle". For example, it is stipulated that if Ethereum falls to $120, you will lose your pledge, so who will define whether Ethereum is $121 or $119 at this time? MakerDAO uses the median value of multiple data sources and transmits prices through the blockchain network. However, in the sharp drop in the market, the entire network is congested, making it impossible to transmit and update market prices in a timely manner.

The second “auction” — that is, the liquidation mechanism that deals with collateral. MakerDAO has multiple problems with the collateral liquidation mechanism:

One is that during the auction, the stable currency Dai issued by the project itself has to be used to buy collateral, but the entire market lacks Dai. Originally, there was not much circulation in the market. Under the panic of the crash, no one was willing to pledge ETH to mint new Dai, and a large number of Dai loans were repaid, resulting in the exhaustion of Dai liquidity in the entire market. The original goal was to be equal to USD 1 :1 exchange, at which point the price rose to $1.12.

The second is the design of the auction mechanism on the blockchain adopted by the liquidation: starting at 0 yuan, submit the auction price on the Ethereum blockchain, and you will win the auction if no one bids a higher price within 10 minutes. At that time, the network was congested, and the transaction fee on Ethereum was more than 10 times the normal one. Users who pay normal fees to participate in the auction cannot submit their prices in a timely manner.

So someone paid a relatively high fee, successfully submitted a 0Dai auction price, and bought nearly 4 million U.S. dollars in ETH, and the entire system fell into an insolvency situation.

Third, "pledge" - whether this "bank" can guarantee the safe custody of the pledge you put there. MakerDAO did not have any problems this time, but what we will introduce below is the attack on this.

April: $25 million stolen from Lendf Me

On April 19th, the decentralized lending project Lendf.me was attacked by hackers, and nearly 25 million US dollars of assets in the project were taken away. Of course, very dramatically, the hackers returned all the money later.

How did you do it? To put it simply, it starts with the accounting design of Lenf.me.

When you come to this bank to withdraw and deposit money, your passbook registration method is set up in the system first. At the deposit counter, the calculation method of your account balance is: (the numbers in the circle are examples).

At the cash withdrawal counter, the account balance calculation method is:

But for example, saving money, everyone thinks that it is best to intersperse some other operations. For example, before entering the account, go to the anti-money laundering center to check whether the source address of the money is from a robber who robbed the bank, and it must be banned first. So Lenf.me adopts an improved standard on the Internet, which supports the function of calling external operations on the basis of compatibility with the original logic.

This function was targeted.

Because the operation called by the hacker is to withdraw money.

Let's solve a problem:

If your account starts with 5 yuan, and you go to the deposit counter to deposit 1 yuan. During the process of depositing money, you go to the cash withdrawal counter and withdraw 2 yuan. What should be the final balance?

As a result, Lenf.me registered 6 at the end

We know the answer should be 4, but during the whole process, although the cash withdrawal counter registered the withdrawal action, the final confirmation was made by the deposit counter, and it was only told that the calculation method is the balance when you deposit + deposit The money you put in, at first glance, 5+1=6, that’s right.

There is no problem with the two separate processing methods of depositing and withdrawing money. If there is no logic to prevent maliciousness, there will be problems. The hacker uses this call to continuously withdraw the assets of the bank until all the money is emptied.

What is especially regrettable is that this method of hackers is called "reentrancy attack", which happened in Ethereum a long time ago. Even the day before the incident, everyone had seen the bug and alerted the team, but unfortunately it was not handled well.

Of course, the hacker returned the money in a very dramatic way, and added a "better future" message to everyone. But it's not because of any particularly fantastic magic, it's just that the identity of the hacker was leaked.

What should DeFi learn from traditional finance?

Despite the rapid development of decentralized finance, it is still in its early stages. The above few examples can actually expose some current problems. What should we learn from traditional finance?

First, respect and learn from financial logic.

Finance has developed for thousands of years, and technology and bookkeeping methods are evolving, but many essential logics have not changed. Blockchain has been recognized from the circle of technical geeks, but to a certain extent, what kind of logic exists in the industry before calling out to "subvert" the vertical financial industry? What risks should be considered? When real money is involved, it needs to be treated with more caution.

Traditional finance has accumulated many years of experience, including: (1) select credible data sources as the price source, and make adjustments according to the rules to reflect market changes (2) adopt VWAP, that is, use the volume-weighted average price as a reference (3) establish and improve Mechanisms for verification and emergency handling. Including the vertical time, what is the difference between this price and the previous price? Horizontally, how does one data source deviate from other data sources...etc.

Another point that has been overly ignored is the "liquidation" mechanism, which actually directly affects the pledge rate required for loans. When the price of collateral drops sharply, you can quickly cash out the collateral at the best price, so you don’t need excessive collateral—this is actually what DeFi is currently criticized for.

(1) Traditional finance provides leverage, and the key is a complete set of collateral clearing system: connect the most liquid traders in the world, adopt intelligent trading strategies, including optimal volume and price selection, large amount splitting Orders, real-time monitoring of trading volume and flow control, etc. In order to take into account risks and user experience, some institutions also launch operations such as step-by-step liquidation.

(2) In order to prevent the pledge from being overwhelmed and insolvency in extreme cases, we can also learn from the practices of many centralized exchanges. For example, in normal daily operations, it is necessary to calculate the VaR value of the asset (the value at risk, which is related to the volatility of the asset and the maximum loss that can be tolerated), and conduct stress testing (assuming what will happen in extreme market conditions) etc.) In turn, guide the product pledge rate, the adjustment rules of the total risk exposure, etc.; prepare an insurance fund for liquidation; set the auction process for liquidation, etc.

Second, make a good trade-off between decentralization and centralization.

It is not mandatory to solve all problems through decentralization. For example, if a company issues bonds through the blockchain, the blockchain cannot guarantee that the company will repay the principal and interest when it expires, but we don't need to force all the "trust" problems to be solved at one time, and it will be automatically executed through the blockchain smart contract. Fund assets are settled across multiple institutions in real time, and bonds can guarantee the authenticity of assets when they are resold, which has already brought value to the financial system.

Another example is the oracle just mentioned, that is, the price index. Many people in the community want to pursue complete decentralization, but there are very strong requirements for the frequency of price updates in finance, not to mention the FIX agreement between financial institutions. This needs to be at the millisecond level, and in many cases, the insistence on absolute decentralization may actually bring losses to users.

The key parts of the project should be decentralized as much as possible. For example, we will talk about the biggest advantage of decentralized projects later. Although you are a small organization, through the blockchain, users don't have to worry about you misappropriating my assets at will.

Now there are many small decentralized financial projects. Although the code is open and transparent, there will be administrator privileges on top of it. The administrator can upgrade the overall code, stop it in an emergency, and so on. Many of them are necessary sexual.

However, many projects, including the administrator rights of the aforementioned PoolTogether design, can take away customer assets. This is a design that should be avoided as much as possible. Users should keep their eyes open and pay attention, and once it happens, it will inevitably have a huge impact on the reputation and development of the entire DeFi.

Passing these tests is in turn to set the overall risk factor and what action should be taken when. When there is a problem with the liquidation, what kind of auction process will be carried out, what kind of fees will be charged during the auction process, or how much the founding members will have to spend together as the entire exchange margin operation.

Third, liquidity is king.

In essence, the blockchain network needs many participants to demonstrate its value, but the root cause of many problems in decentralized finance at present is precisely the high barriers to use and insufficient liquidity. This is of course due to the small scale of DeFi, but there are still some things that can be worked hard.

On the other hand, especially in the early stages of the project, DeFi project parties can actively invite and encourage some liquidity providers to join the platform. There will always be a chicken-and-egg problem in building a platform. Only by taking the initiative to solve the problem of relatively concentrated incentives will more providers come to the platform to attract more users.

Although there are many shortcomings, there is still a long way to go before large-scale application, but at the same time we can also see that decentralized finance has played many new tricks that traditional finance cannot play today.

First of all, what we can see is that compared with today's services, decentralized finance has excellent characteristics, especially including openness, standards, transparency, and programmability.

However, many people actually confuse the concepts.

These characteristics are not essential to decentralization. If it can only involve a single subject, today's financial institutions can do it too, they just gave up this road out of commercial considerations.

This is like the birth of Yu'e Bao. In fact, could commercial banks at that time provide more convenient and easy-to-use monetary funds with higher interest rates? absolutely okay. But it is obvious that you can absorb a large amount of deposits by paying low interest rates, so why bother to find trouble yourself.

It has to be said that in fact, using blockchain in many scenarios is a pseudo-demand. Many places assume that there is a strong centralized organization, and the blockchain is not necessarily the most efficient method.

So, what is the difference between blockchain finance and today's traditional finance?

First, on the blockchain network, there is native money, which can improve settlement efficiency across subjects and reconstruct governance trust.

Today on the Internet, the information layer and the capital layer are separated. For example, if you send money from China to the United States, the information can arrive in real time, but the funds need to go slowly or even a few days.

why? Now remittances do not need to actually move gold.

Today the Internet transmits information, and the settlement of assets is the transfer of money, which needs to be registered with different entities: through your local bank, local clearing network, intermediary bank, local bank in the United States...

The blockchain just plays the essence of coordinating multiple subjects to keep accounts at the same time, and the synchronous flow of funds can be realized on the network.

With an asset layer on the blockchain, decentralized governance can be effectively implemented, thereby building trust.

Governance requires rules and penalties. It is difficult to achieve governance independently in terms of pure information network exchange. For example, if you communicate with a person in the southern hemisphere, such as Brazil, on the Internet, and he deceives you, it is difficult for you to punish him only through today's Internet.

To some extent, why do we trust big institutions such as Google, Amazon, and Tencent? We do not believe that big companies are benevolent, but implicitly say that we believe that the benefits they can gain from these big goals against evildoers are much smaller than the losses caused by their reputation damage and offline law enforcement, so they There is no motivation to do evil easily.

However, the network itself has money, and the situation is different. After all, you can scold someone far away online, but the effect is a bit different from being able to deduct his money directly. This asset on the network can be locked, and the settlement can be completed in real time. When our punishment is carried out effectively, we don't need to rely on this kind of centralized trust and large institutional providers.

In this way, the original centralized provider monopoly services can be split and recombined: imagine your personal data information management, product demand can choose providers at will, and various financial products can be combined with other services at will. Many truly decentralized and rich financial ecology can be grown.

Second, where centralization is absent, the blockchain provides a way to coordinate multi-agent cooperation.

Many people in the encryption circle think that the public chain is the blockchain, and the alliance chain is not everyone's ideal. But in fact, the latter actually coordinates the cooperation of multiple subjects.

In fact, there are many things that need a centralized institution. For example, if different domestic banks want to share the offline merchants of each bank and improve the efficiency of liquidation, UnionPay actually connects the banks.

But even though UnionPay is already considered a large institution supported by the state, it has encountered a lot of resistance in the course of its development history. I will not talk too much about the development history of UnionPay and NetsUnion due to time constraints. It is actually very interesting. You can take a look at the difficulties in the rise of a centralization.

The state cannot penetrate into all aspects, let alone some small scenes, and centralization is actually out of reach.

When centralization cannot be reached, even if cooperation can increase mutual benefits, why do I need to follow your instructions and ask me to pay you as much as you want? I don't want to listen to you, and you don't want to listen to me.

At this time, the blockchain provides a coordinated method. It’s not that I listen to you, and it’s not that you listen to me. Everyone can become a bookkeeper. According to the rules and forms of consensus, everyone forms an alliance—a distributed database, but multiple parties have the right to bookkeeping. In fact, it is Blockchain.

Third, it sounds more metaphysical, but it can actually bring about an impact on the industry. Blockchain finance represents a new mode of thinking, a new way of organization and a new form of service.

As mentioned earlier, Yu’e Bao did not provide the latest services back then, but it was the efforts of financial technology that pushed the entire industry forward.

Another example is at the information level. What is the difference between the content produced by a centralized organization, such as Xinhua News Agency, and the open WeChat official account and Douyin video platform? You said that Xinhua News Agency is stronger in reporting ability and professionalism and ability to mobilize resources? Or is a small WeChat public account/person who uploads videos on Douyin even more powerful?

In terms of overall strength, Xinhua News Agency may be more powerful, but the latter offers a new possibility: it provides an open platform and gives each individual a creative space under simple rules. Maybe Xinhua News Agency is better than a WeChat public platform that writes fashion magazines from the perspective of covering major national events; but in this small aspect, what the fashion commentary public account writes is more in line with the needs of some readers.

The blockchain not only provides such a platform for information, but also for assets. Under simple rules, open finance can be built, which will release more resources and productivity.

To some extent, the idea of ​​blockchain is actually in line with the "complex" system theory that we are gradually discovering in the natural world. In a complex system, each small unit is only given very simple rules, but under the rules, various possibilities can be released, and a huge vitality is radiated.

It represents a new possibility, which will be greater than the current impact of financial technology on the banking industry. It will also inevitably force the reform of the entire industry and redistribute the industry structure during the development process.

Q&A

secondary title

Q1: Will the situation of DeFi projects become normalized? How is the industry responding to this situation?

A1: I think it will be normalized. First, it is a new thing, and the accumulated risk awareness in the industry may not be sufficient. In addition, it is more open and transparent than centralized finance, and it is easier for everyone to see your weaknesses. Someone even mentioned that the blockchain is now a golden 10 years for hackers.

But we think this may be a necessary stage in the development of a financial industry or a new technology. SWIFT has had major attacks until 2015/16, including the central bank of Bangladesh, etc. have been hacked, and many bank risk events may not tell you.

So I think it’s a normal thing to just calm down, and everyone can accumulate more and more experience like this, just like today’s traditional financial risk control, usually after many events in the past. A process that will be perfected.

So we may say that today's DeFi is a pioneer and a test field for the overall large-scale blockchain financial applications in the future.

Q2: In addition to traditional finance and DeFi, will DC/EP have an impact on traditional finance? Will DC/EP and DeFi have a chemical reaction?

A2: First of all, the understanding of DC/EP refers to the DC/EP of the People's Bank of China. Will it have an impact on traditional finance? I think it will. It is the digitization of M0, that is, the digitization of cash, and will also realize dual offline payments. To some extent, the introduction of DC/EP may help banks and some operators gain a place in today's monopoly electronic payment landscape unified by Alipay and WeChat.

Will DC/EP have a chemical reaction with DeFi? At present, as far as the public information is concerned, it has no restrictions on technical routes, and it has not said that blockchain services will be used, but there may be certain blockchain components in this technical attempt, which is not related to decentralization. Finance produces some organic combinations.

Q3: DeFi on BTC seems to be very popular this year. DeFi anchored to BTC such as tBTC and WBTC are rising rapidly. How do you see its future?

A3: First of all, the current ceiling of DeFi projects is basically on Ethereum. In today's cross-chain immature situation, it is its overall network users and the market value of assets on the network.

We can clearly see that the market value of BTC is much larger than that of Ethereum, and the attention it receives is far greater than that of Ethereum. Traditional institutions still look at BTC, after all, its market value is the first.

The rapid rise of BTC-anchored projects is a way to introduce BTC to Ethereum, the main platform for DeFi. For example, pledged loans. If you look at the decentralized financial lending platform, most of its volume comes from BTC pledged loans, not ETH pledged loans. Its market value is there, and your business volume Limited by the value of the assets being pledged. Then the larger the amount will naturally promote you.

You can think of this matter as issuing an asset on Ethereum to anchor another asset in the real world. For example, USDC, these US dollar stablecoins, are also anchored to another asset - the US dollar in the real world. Like WBTC, they find a centralized and relatively trusted platform for hosting.

WBTC is to put BTC into BitGo (Note: a licensed custodian institution located in New York State, USA) to issue "BTC stable currency", an asset denominated in BTC, which increases the richness of the entire ecology. Of course, tBTC does it in a decentralized way. The richer the assets on the platform, the more financial services you can build, so it will definitely become a good driving force for the development of the entire DeFi.

Q4: What is your opinion on the skyrocketing market value of stablecoins in the past two months?

A4: There are many reasons. One is that after March 12, the market de-leveraged and some digital currencies were replaced by stable coins. It is a bit like many traditional investors sold stocks and other risky assets and held stable dollar assets. It is one; the other point is actually a good thing, there are some traditional institutions entering the market. Of course, it is very important that Tether, an organization that is not particularly reassuring, has recently issued a lot of USDT. There are different opinions on the reasons, so I will not comment here.

The time is almost up, so this is the end of our sharing today. Welcome everyone to have the opportunity to communicate more, thank you.

Contributions in this issue｜Rong Lin