Three top scholars of global blockchain economics talk about POS, Staking, DAO

Published on 9/27/2022   198 views   0 Comments  

On September 26, at the 8th Global Blockchain Summit - the Great "Star" River Block Chain Economics Theme Forum held by Wanxiang Blockchain Laboratory, Lin Zhiguo, a chair professor of management at Cornell University and a tenured professor of finance at the National Institute of Economics of the United States, and He Zhiguo, a professor of finance at the University of Chicago Tang Ke, the director of the Institute of Social Sciences of Tsinghua University, and Zou Chuanwei, the chief economist of Wanxiang Blockchain, had an in-depth discussion on the frontier issues of blockchain economics. The following are the main topics discussed by the guests (with deletions):



Zou Chuanwei (host):

The three professors represent the cutting-edge level of research on blockchain related economic issues around the world, and there is also a lot of cooperation between the three professors.

In his keynote speech, Professor He discussed the economics of PoS in depth, and took Ethereum beacon chain as an example to illustrate his concern about the centralization of the Staging pool. In the PoW blockchain represented by Bitcoin, the phenomenon of mine pool centralization has long appeared. Excuse me, Professor He, do you think PoS blockchain will encounter this problem sooner or later? Is it possible to alleviate these problems through consensus algorithm level design? This problem is of great urgency at present, because Ethereum has switched to PoS since September 15.

He Zhiguo:The question is "one bill may one trillion question". This problem has two levels:

First, will PoS encounter the problem of centralization sooner or later? The original design should be decentralized, but it will gradually develop into a centralized and centralized trend. From the current data, this trend will change earlier. Because at the beginning, PoWs were "small play", without professional teams or very professional players. Later, they gradually found ways to benefit, and then there was a scale effect, which would lead to a centralized trend. Now, PoS will soon show such a trend, which is a very simple economic principle.

This problem has nothing to do with Crypto or anything else. It is the most basic problem of scale effect of economic development. From this point of view, it is not the mechanism design at the algorithm level that can solve this problem. Of course, if there is a higher level of mechanism design, it can be solved, such as tax collection.

Of course, things are changing slowly. Blockchain can make it easy to implement a very effective tax mechanism at the social level. From this point of view, I don't worry about too much centralization, but at the beginning, everyone will see centralization immediately.

Zou Chuanwei (host)Thank you Professor He. Professor He mentioned that it is not only an algorithm problem, but also comes from the professionalism of the Staking pool and the inherent scale effect in economics. This question is just related to the question I want to ask Professor Cong. In Web3.0 economics and the three promising directions, Professor Cong mentioned the relationship between Web3 identity and reputation, especially the complementary and alternative relationship between Stack and reputation.

Excuse me, Professor Cong, do you think there are more complementary elements or substitutive elements between Stack and reputation? Especially in the current centralized and professional trend of the Staking pool, if I am a small white user who has just entered, I have very few stakes, but my reputation is very good. How likely do you think I will rely on my reputation to counter attack in the market? If there is no possibility of counter attack, the degree of solidification of this market will be very obvious, unless we go to the tax method just mentioned by Mr. He.

jungle:All your questions were excellent. Most blockchain mechanism design and consensus design, regardless of computer technology implementation, are indeed economic problems in many cases. Most of the incentive mechanisms are basically static ones, such as Proof of Strike and Proof of Work, which generate a block each round and then provide corresponding incentives. "Reputation" is still a relatively new concept. Although it is not strange in the traditional economy, it is still a new direction for blockchain, Web3 and DeFi. It can be complementary or substitutive with Stack, depending on the specific application.

Returning to the topic just discussed - the problem of centralization, "winner takes all" often appears in traditional economic platforms and digital platforms, including Taobao and JD merchants, who also have ratings, user ratings, good comments, and high scores, so they may get more business, and also have more opportunities to build reputation and develop business better. This phenomenon also exists in many online live broadcast and cargo economy, and there is a strengthened feedback mechanism.

At this time, Staging has a way to break the concentration problem in the traditional economy. The nodes may not have established a good rating, but if more Stages are available, they may also be very competitive. Because more stakes face greater punishment, this solves the cold start problem in the traditional economy.

Back to blockchain or blockchain related applications, I think there are more complementary components. Why? Because the design concept of traditional blockchain is static incentive mechanism. Dynamic incentive mechanism can provide incentives more effectively. It is called "Back Loading Incentives" in English, which can push incentives to a further future and make them more effective.

I worked with ChainLink to discuss the problem of oracle network design. There are some so-called "impossible theorems". Sometimes it is difficult for all Oracle oracle machines to provide correct and accurate information services. However, if the game is repeated, it will last for a long time. At the same time, it can provide a dynamic incentive mechanism. Through reputation building, all oracle machines can better complete their tasks.

There are many such examples, including in some cyclical markets. When the market is weak, if only dynamic mechanisms are used, it is not enough for participants to complete their tasks well. However, if reputation is combined with Stack, it may help the system survive a relatively low cycle. When the market recovers, many participants have established a good reputation, so they will actively maintain the system as "Incentive Compatible".

Of course, the specific problems still need to be analyzed specifically, but the general direction is determined. If the reputation system can be added to the original static incentive design, it may be very helpful.

Zou Chuanwei (host)Thank you for your reply. Professor Cong introduced a very cutting-edge problem in the current mechanism design - the combination of dynamic and static mechanisms. I believe this is also a new problem in economics, and there will be many innovations in mechanism design.

The next question is for Professor Tang. Professor Tang's empirical analysis of the staking economy has inspired me. I also thought about verifying whether interest rate parity is valid in staking. After reading Professor Tang's article, I think my confusion has been basically solved. Professor Tang found that the higher the rate of return of staking, the higher the proportion of users staking, and the greater the motivation for token appreciation. However, another phenomenon is also found in the industry, because many times Staging is accompanied by the issuance of new Tokens. If the higher the yield of Staging is, the higher the issuance speed is, which will also cause the higher the number of Tokens available for sale, which will exert downward pressure on the price. From the perspective of Mr. Tang's article, these two influences should be the former one. However, do you think there is a concept of optimal staking yield, and the higher the staking yield, the better?

Tang Ke:Thank you. This is a particularly good question. When we consider the empirical study, we first assume the exogenous given mortgage return rate. Under the exogenous given mortgage rate of return, we can see the positive proportional relationship you just mentioned with the existing sample - the higher the rate of return, the more mortgages. When the designer sets the mortgage guarantee rate to a very high level, the positive relationship is not obvious. When the exogenous rate of return is set, investors and mortgagers are rational. They will consider whether such a high rate of return is consistent with reality, or whether such a high design can sustain the platform.

To return to your question, under given conditions, or in the sample of existing rational investors, this relationship is obvious, and in the sample, very high returns are rare. If we only look at the high rate of return, there is no positive relationship or it is not obvious.

In addition, is there an optimal point? I think if the return rate is regarded as an endogenous variable, it can be adjusted. In this sense, the platform stability should have the most advantages according to the different use contents and scenarios of the platform. But at present, the optimal mechanism is often difficult to exist. Usually, the return mechanism has been clearly written in the initial investment prospectus, and it is difficult to deviate from this mechanism later.

Therefore, it exists in theory, but it is not easy to realize in practice. Maybe there will be a better way to adjust the rate of return more endogenously in the future, which may lead to a better mechanism design method.

Zou Chuanwei (host)Thank you for your answer. In theory, Mr. Tang believes that the optimal Staking yield should exist, but in reality, there are many challenges to find an appropriate balance in various problems and scenarios, which also need to be further observed and tested in industry practice.

Next, according to the keynote speeches of the three professors, I designed some questions for the three professors to listen to your opinions.

First, Professor He spent more time in his keynote speech to elaborate on the functions of smart contracts. On September 8, Fan Yifei, Vice President of the People's Bank of China, introduced the programmable application of digital RMB smart contracts at the "Second China Beijing Digital Finance Forum". In the development of decentralized financial DeFi, everyone felt the impact of the composability of smart contracts. Because of the composability of smart contracts, DeFi was referred to as "Lego building blocks". As far as I can see, there is no concept corresponding to programmability and composability in economics at present.

From the perspective of economics, how do the three professors view the programmability and composability of smart contracts? What is the impact if it is used in mechanism design? Will it cause new risks?

He Zhiguo:These are two cross-border concepts, which actually tell two things.

First, citing the application of China's digital RMB smart contract proposed by Fan Yifei, under the almost centralized approach, it is very easy to do digital RMB, which is a very direct technology application. However, in terms of decentralization, there is still a big difference compared with the smart contract thought in general economics, because the executable input of the smart contract must be on the chain. At present, I rarely see cases where on chain and off chain are well combined. Because the things under the chain have not been put on the chain yet, I don't think we can just discuss smart contracts and ignore the most important issues.

Second, programmability and composability. For economists, there are not many ideas on this issue. In many cases, contracts of different financial institutions may be borrowed from each other, which is similar to all contracts of lawyers. The most important points and figures are left behind. The same is true for personal salary contracts. All languages are the same, but the most important figures are different. This is actually the problem of programmability and composability.

When it comes to risk, think of all parts as "Lego bricks" one by one. If people who make investment strategies do not consider the consequences of different combinations, the risk will be great. At the end of the application, we will definitely try different combinations internally, and what will happen.

How to use smart contracts in real life is more traditional economic and financial thinking, which abstracts "building blocks" in all aspects. Think about what interactions are there, and then make judgments.

jungle:Thank you Professor He. He spoke very well. Let me add a little.

First of all, I think there is little to say about programmability from the perspective of economics, computer science and cryptography. One of the reasons, as Mr. He has said, is the problem of off chain enforceability. The issue of execution is not considered in economics, because the contract theory is generally discussed, and it is assumed that the contract will be executed. From this point of view, the two sides have little to learn from and complement each other.

From the perspective of composability, economics still has a lot to learn. Most of the contracts we see in our life and research are two-way contracts that need to be signed by both parties. But there are also some contracts studied in economics that have a combination nature. For example, a contract stipulates that my remuneration depends on my competitors. At this time, the remuneration I get and the incentive mechanism I face will interact with the contracts of my competitors and partners. It can be imagined that a large investment institution will invest in many small funds, and they will make a permutation portfolio. Although each of them will sign a contract separately, they actually have a comprehensive consideration.

In this direction, economic research and experience can be used for reference for smart contracts. Or conversely, if the smart contract is well implemented, it can also inspire more theoretical research on economic contracts.

Zou Chuanwei (host)Thank you both. Professor He and Professor Cong discussed this issue from the perspectives of centralization, decentralization and different scenarios on and off the chain.

The three professors have conducted a very in-depth study on Staking. We can see two scenarios: the first scenario, in the PoS blockchain, Staking guarantees the security of consensus algorithm; The second scenario is that in the DeFi stage, Staking guarantees the debt performance in the decentralized environment. On the other hand, it is also necessary to see that Staking will undoubtedly reduce the efficiency of liquidity use. For example, DeFi's development indicator is TVL's total lock position value, which has positive value, but in turn, it also shows that many assets are locked.

Is there an optimized Staking design that can not only guarantee the consensus algorithm, security, and DeFi debt performance ability, but also reduce the occupation of liquidity? Some time ago, Vitalik proposed the concept of SBT soul binding token, trying to introduce reputation and mechanism into a decentralized environment. Do you think Vitalik's SBT is a feasible solution?

jungle:Let me share my thoughts on this issue, mainly in three aspects:

First, discuss the impact of the design of Staging. It should not be considered as an independent thing. It should be combined with the whole system, especially the system using tokens. That is to say, when Staging, there are many different Tokens in the Token or Lock up. At the same time, it should be combined with the monetary policy in the system. The so-called "lock" reduces the liquidity, but also affects the number of movable tokens. Therefore, a joint design of Point design is required. The design of Staking should be considered not only from the micro level and corporate finance level, but also from a more macro perspective.

Second, when considering the Staking optimization scheme, we should also consider the stage of the system. Teacher Tang has mentioned before that it is difficult to design an optimal interest rate in practical operation. Many projects will set higher interest rates. Of course, you can say that this is a risk compensation. This is just like Amazon, Uber and Didi. At the initial stage, they will give users a lot of subsidies to attract users to perform Staging at this stage. The mode of staking reward mechanism may be different from that of long-term stability. After the determination stage, we can determine many economic studies to reduce liquidity, pledge, deposit, interest rate in the economy, or real estate investment.

Finally, return to SBT. SBT is actually grafting "reputation", or more broadly, it actually points out the core concept in Web3, namely "Identity". The original Identity is limited to one platform, and your information and behavior on this platform cannot be universal across platforms. Is there a cross platform "Identity" that can be used even on and off the chain? I think the direction and concept of SBT is correct. Adding reputation and combining it with a static incentive mechanism will have many benefits.

However, the details of the implementation of this concept return to how to design reputation mechanisms. Reputation or Rating mechanisms may involve on chain information and off chain information. Who will do it and how fast it will be updated are all dynamic optimization problems. One of the links is related to Staging. There have been some studies on how to optimize the design when combining reputation and Staging, and this issue is also a direction worthy of in-depth study.

Tang Ke:My idea is also related to Staking. From an empirical perspective, when there are more stakes, the liquidity of the whole market will decline, and the Token price may increase.

When we keep accounts and mine on the platform, the traditional way is to pledge, which is equivalent to the collateral or the investment form. I tell you that I will not commit evil, because I pledged so many things. If I commit evil, the platform will have problems and I will be unlucky. What if we use SBT's reputation mechanism? As Mr. Cong said, it can help us solve the problem of the quantity of collateral, or it may be that the two will eventually be combined. Some people have a very good reputation, so they can keep accounts without so much collateral.

Therefore, I believe that the combination of the two is most likely to be realized. For example, if we borrow money in the bank, we need to provide collateral to the bank. However, if the borrower's credit is particularly good, there will not be so much collateral, or even collateral.

He Zhiguo:After reading the SBT paper, I realized that this thing was "sesame credit". The idea of SBT is a very simple and direct economic logic in real life, which is actually an incentive mechanism. It's a good idea, but I think there must be many difficulties in implementing it. The reason is that you want to be private and want to do things that no one knows, but finally you want to go back to the basics. I hope everyone can see what you do and need others to give you a score. I don't know how to achieve this.

Zou Chuanwei (host)Thank you Professor He. The answers of the three professors provide a very comprehensive perspective for us to understand these questions. Professor Cong and Professor Tang mentioned that reputation should reduce the use of collateral to a certain extent. Teacher He mentioned that he had hoped to participate in the market in a relatively anonymous way, and did not want too much privacy to be recorded, but finally he needed someone else to help rating. This was a dilemma. These perspectives are very helpful for us to understand the following applications.

Professor Cong mentioned the relationship between Token destruction and monetary policy in his keynote speech. Token destruction is one of the very common mechanism designs in the blockchain field. The question is how do we understand token destruction from an economic perspective? Is the destruction of Token closer to the repurchase of base currency by the central bank or closer to the repurchase of shares by listed companies? Or just to create scarcity and special value? Although there is a lot of practice on token destruction in the market, three professors, do you have an optimal mechanism design for token destruction?

jungle:I would like to share some conclusions from my previous research with the other two professors. We should not directly compare the destruction of Tokens with policies such as quantitative easing and tightening. When Tokens are destroyed, they are completely destroyed. It is not a kind of exchange, nor is it a way for me to issue new coins and then buy them back. The reverse is also true. In practice, Token destruction is closer to the destruction of the base currency. But conceptually, it is better to adapt to the cycle of economic growth by adjusting supply. Many luxury brands, such as Burberry, burn some goods when they have too much inventory, which seems wasteful, but it is for the sake of long-term scarcity value. "Scarcity" is an artificial design, which can be optimized according to economic principles. However, many platforms do not make clear what they want to optimize. What is the purpose of the monetary supply and destruction policy on the platform? The Central Bank made it clear that in order to ensure the employment rate and the stability of prices.

What should we pay attention to in the ecosystem of blockchain and Web3? Pay attention to the stability of the price? Or is the value of its exchange with other currencies stable? Or let it grow as much as possible and have the highest applicability for everyone to use? At the same time, we know that there are also security problems in these systems. Can this money supply policy provide a good enough and comprehensive incentive mechanism? I think the best design can be obtained after these purposes are clarified. The optimal design is not necessarily complete destruction. If you can print and issue new Tokens at any time while completely destroying them, then complete destruction is not complete, just like the company can still sell the shares after buying them back.

He Zhiguo:Between Token destruction and monetary policy, there is a small economy and a large economy. I agree with Miss Cong's idea. I don't think Token destruction itself has too much meaning. Token destruction may have an immediate effect, but I'm not sure. As Mr. Cong said, it is not very meaningful to destroy today and add tomorrow.

Zou Chuanwei (host)Thank you Professor He. Professor Cong, how do you think the data DAO can help users better protect their data rights through the collective bargaining mechanism. There has been a lot of discussion on data trust in Europe and the United States recently. Do you think the data trust they are discussing is a special form of data DAO.

jungle:This is also an issue that I have paid special attention to recently. Blockchain, financial technology, and Web3 have brought a lot of financial benefits. When users participate in and contribute a lot of data, have they been well rewarded? Or is his privacy well protected? There are many policies and discussions about this, such as privacy protection policies, GDPR (European Union General Data Protection Regulation), CCPA (California Consumer Information Act), etc., as well as the Open Banking, Open Data Initiative and data sharing policies that Teacher He is working on.

Is Data Dao, or user alliance, or data DAO a good solution? First of all, think about what the problem is. The problem is that users are fragmented, and the information and data they generate are also fragmented. They will not take their exogenicity into account when making decisions. Maybe I would like to share my data if the platform gives me some benefits, but I have not considered whether the data I contribute makes the platform unique or affects more competition in the market? From the perspective of long-term development, will more users suffer more harm instead. This is similar to antitrust.

In legal research, people often talk about Data Trust. Data DAO and Data Trust are not exactly the same. t. In real life, there are also some examples that have implemented Data Trust, but most of them are still handled by intermediaries. Data DAO itself is a decentralized management mode. Users can choose some representatives and platforms to "bargain" with each other, and then distribute them to scattered users as the remuneration of the overall platform.

At present, there are some similar attempts, but no particularly perfect cases have been seen. What we are studying now is how to design a mechanism to better coordinate the decisions of all users, so that the "Consumer Welfare" can reach the optimal state.

Zou Chuanwei (host): Thank you, Professor Cong, and look forward to your next research progress in Data DAO.

Next, I want to discuss a trend question. The three professors are experts in Token economics. What do you think of the research field of Token economics? Which research fields in economics, such as mechanism design and monetary economy, intersect with it? What are the new problems? What analytical tools do we need to introduce now?

Tang Ke:Token Economics has some new tools. I think the most important tool is public key cryptography. The whole blockchain is built on two very important tools:

First, public key cryptography. The entire account system is derived from RSA, Diffie Hellman and other public key cryptography.

Second, the establishment of consensus mechanism. I think Nakamoto Cong's great contribution is the PoW mechanism design, which makes people know that consensus can be reached through this method.

For public key cryptography, users can leave the original centralized account by designing a complete set of public key cipher modes. The public key cryptosystem can understand many deeper problems, such as the Turing completeness problem and Ethereum problem mentioned earlier. In fact, most of them analyze this field from the perspective of cryptography.

At the same time, what research fields do they intersect with? Platform economy, bilateral market design, and some new assets emerging on the platform. Everyone is debating whether such assets are commodities or currencies. I think these assets are different from traditional commodities. What are the differences? How should we analyze these virtual assets? There are also some new extensions of traditional asset pricing theories, such as what the new influencing factors of cryptocurrency are, and whether the traditional network influencing factors and traditional stock market factors still exist. These are new tools.

He Zhiguo:I add two more that I find interesting. First, competition. With Token, the platform gives users a portion of its pricing behavior, which is an interesting phenomenon. Many politicians, including the United States and China, are concerned about the issue of platform oligarchs. I believe there will be more and more research in this area in the future. The second is information economics. Information Economics and the public key that Mr. Tang said are very interesting, but the application has not reached that level, and there is no good interface with information economics considered by ordinary economists, but it does have direct access to information, including Zero Knowledge Proof. In the article I wrote with Jungle at the beginning, it was actually emphasized that this was about information, but I hope it can have more practical applications. Information is an important pillar of economics and can be reflected in future blockchain applications.

jungle:To add simply, Mr. Tang and Mr. He mentioned earlier that this is related to monetary finance, corporate finance, asset pricing and market microstructure, including how to design the transaction costs in the system.

From the perspective of practitioners, some things from economics and statistics may be helpful. There are two points worth emphasizing: first, we will see a lot of open data on the chain and also many networks, so the work of blockchain tracking and detective can mine more information through the big data analysis model, and will have more understanding of the whole network; On the other hand, in the Web3 environment, we may not achieve complete decentralization in many cases, but are still distributed in many cases. In this case, it is difficult for traditional analysis tools, such as Closed form solution, to completely solve some models or completely simulate some scenes.

Some other tools may be better used, such as Agent Base Simulation. You need to simulate many different types of participants on a large scale, and then see what the macro effect is. Or as Teacher Tang and I mentioned in a study, many individuals are solving different problems. How to describe macro phenomena? Maybe this kind of tool will have more applications.


Generic placeholder image
215 views   0 Comments   1 Weeks ago
94 views   0 Comments   1 Weeks ago
87 views   0 Comments   1 Weeks ago
128 views   0 Comments   1 Weeks ago
114 views   0 Comments   2 Weeks ago