Hello everyone! Today we are very honored to have Prof. Campbell Harvey here as our guest speaker. He will talk about “DeFi and the Future of Finance.” Let me introduce Prof. Campbell Harvey First. Prof. Campbell Harvey is a Professor of Finance at the Fuqua School of Business, Duke University, and a Research Associate of the National Bureau of Economic Research. He served as President of the American Finance Association in 2016, and he has been named “Quant of the Year” for 2020 by the Journal of Portfolio Management. He edited The Journal of Finance — the leading scientific journal in his field and one of the premier journals in the economic profession from 2006–2012. Prof. Harvey obtained his doctorate at the University of Chicago in business finance.
Even though he has so many titles and achievements, the thing that impresses me the most is that he is always a self-disruptor and a pioneering scholar for his lifetime.Over the past six years, Professor Harvey has taught Innovation and Cryptoventures at Duke University. The course focuses on blockchain technology covering both the mechanics of blockchains as well as practical applications, and it is extremely popular among Fuqua students. He also teaches Tech Driven Transformation of Business and International Finance. Prof. Harvey also offers a Coursera course “Blockchain Business Models”, which enjoys a fantastic reputation. His recent publication, DeFi and the Future of Finance, is available for pre-order on Barnes & Noble and Amazon and has drawn wide attention from researchers and finance practitioners. In the fall, he will launch a four-course Coursera sequence under the title of DeFi and the Future of Finance. The courses are DeFi Infrastructure, DeFi Primitives, DeFi Deep Dive, and DeFi Risks and Opportunities.
We just cannot wait! Now, let’s welcome Prof. Campbell Harvey for his inspiring lecture.
Prof. Campbell Harvey’s insightful conclusion as a call for action to the young scholars:
This is from a computer scientist that I respect a lot, Tim Roughgarden’s first open lecture of his course Foundations of Blockchains in computers science at the university of Columbia. And I want you to consider this:
The innovation draws on parts of computers science (e.g., cryptography and distributed systems) and other fields such as game theory and finance and is developing into a fundamental and interdisciplinary area of science and engineering its own rights.
Future generations will be jealous of your opportunity to get in on the ground floor of this new area — analogous to getting into the internet and the Web in the early 1990s.
I cannot overstate the opportunities available to someone who masters this material — current demand is much, much bigger than supply.
I want you to consider this: “Future generations will be jealous of your opportunity to get in on the ground floor of this new area.” OK, just reflect on that. You are in a great position where less than 1% into this disruption. You get the opportunity to get in on the ground floor.
We are now seeing the scaffolding of a new city that reinvents finance. It is just a matter of time for the legacy players — and they know it.
Millenia ago, we started out with peer-to-peer market exchange — barter. We have come full circle. All assets, physical and virtual will be tokenized.
Open AMA with pioneering students
1. DeFi risk in Flash-Loans
Tianyu Wu (Student):
In your book DeFi and the Future of Finance, you mentioned that one of the emerging DeFi applications, flash loans, could highly reduce the risk of default when borrowing or lending on a decentralized platform. However, a potential risk in the smart contract risk might allow users to take advantage of arbitrage opportunities. Will you think of the risk as an illegal frontrunning behavior, or an acceptable arbitrage behavior?
Prof. Campbell Harvey:
I think that it is important to distinguish between hacking and arbitrage. All of these data are available, the code is available, you see an opportunity, and that's an arbitrage opportunity. And if there's a weakness in a contract, you can exploit it, it's nothing against the law, you can also front-run. So in Traditional Finance, front-running a trade is illegal, and you will go to jail if you actually do that. And there was a recent case where a customer put in a big order to buy some stock and then the broker actually goes out and buys some call options on the stock before the order goes through. That's illegal. In this space, front-running is completely legal because all of the information is public, and anybody can see the pending orders in the memory pool. So here it is a cost of this system, but it is definitely not illegal.
Thank you, Professor Harvey, and your comments really bring me up. When it comes to adapting the original concepts in traditional finance to the new crypto world, we should be aware of the different contexts for sure. In this example, we should be clear about the public or non-public information, which is similar to that about the concept of public information or insider information in the traditional stock market. But I am still wondering whether the insider information still exists and only the miners can be accessible to and make use of this information since only this group of people have the mining power. If so, I am curious whether this potential value is of great extra benefit to the miners, and whether this phenomenon would do great harm to the overall blockchain security.
Interviewee bio:
Tianyu Wu, a junior Math student at Duke Kunshan University (DKU), full admission scholarship recipient, and 2019-2020 unique National Scholarship recipient. He has a solid foundation in Mathematics and excels in disciplinary courses. His areas of interest include quantitative analysis in the financial market, blockchain and cryptocurrency, and game theory. He is now doing cutting-edge research on Automated trading algorithms in FinTech. This research is sponsored by the Summer Research Scholars Program. He is now working under his advisor, Prof Luyao Zhang, with the topic of front-running in the crypto world. At the same time, he was also one of the E-Board members of SciEcon CIC, leading the development of the HR management system in the era of transformation.
2. Metaverse
Xinyu Tian (student):
With the development of Augmented Reality (AR) technology, the concept of Metaverse has absorbed people’s attention nowadays. The CEO of Facebook, Mr. Mark Zuckerberg described the Metaverse as a world with a maximalist, interconnected set of experiences straight out of SciFi. And Metaverse will definitely be an interactive space for DeFi. What are the new opportunities and risks for DeFi in the Metaverse?
Prof. Campbell Harvey:
This question is about the possibility of a different world, or the so-called Metaverse, where we’re interacting essentially virtually. I actually believe in this vision that in the massive upside of linking decentralized finance to the gaming world. So this has only just begun. With Non-Fungible Token especially, I can see a lot of upsides here in terms of somebody that’s really thoughtful in the gaming space and merging that with decentralized finance.
Thank you so much, Prof. Campbell Harvey! With the development of high technology, I do believe that the idea of Metaverse will come true one day and people would be able to utilize the advanced technology to support their life. And maybe at that time, the geographical quarantine would never separate us and people can reunite virtually. DeFi and NFTs would be the mainstream of world finance and people are living with trust.
Interviewee bio:
Xinyu Tian is a junior student majoring in Data Science at Duke Kunshan University. She is a full admission scholarship recipient and a researcher at the Environmental Research Center. She is passionate about the interdisciplinary study of Machine Learning and Intelligent Economics, as well as the disciplinary study of deep learning. And she is currently working on game theory problems and consensus mechanism design in blockchains, supervised by Prof. Luyao Zhang. Besides, she is also the Director of Communication at SciEcon CIC, working on SciEcon publications and assisting the AMA program.
3. DeFi Risk Measurement
Haoxin Yu (Student):
The last chapter of your book DeFi and the future of finance illustrates DeFi risk such as smart contract risk and governance risk intuitively. It inspired me to conduct relative research on measuring these risks. I am wondering which of these risks you elaborate on can be quantified?
Prof. Campbell Harvey:
There are two things to illustrate. The first one is that we are less than 1% into this disruption and it’s still very clunky to deal with decentralized finance. It’s awkward that the user interfaces are not ideal, and it’s been a while before people interact in a way that they don’t need to understand what’s going on, they only simply interact with these protocols. I regard this as a risk of adoption. The second one is obviously the regulatory risk, which is massive. I think the other risks can be largely mitigated in the future, but there will be a time where the regulators are fighting for control. They consider it a bad thing for the central bank to lose control. Let’s take Venezuela as an example. Venezuela is a country in hyperinflation, if you’re a rich Venezuelan, then you’ve got a bank account and US dollars in Miami. For everybody else, they get hammered. But today, for everybody else, each of them can get a mobile phone. They can store a token on the mobile phone that is backed by US dollars. As a result, they didn't need a bank in Miami for US dollars and got their own “bank”, which would put the central bank out of business. To some extent, they should be out of business because they introduced a reckless monetary policy that generates 700% inflation. I don’t think it's a big problem to switch to alternative ways to exchange value, we think of a common currency. However, in the future, everything will be tokenized. There’s no problem if you want to pay in gold with a token and you can also return to the gold standard. For example, you can do this by getting a token backed by gold, by real estate, or by equities today.
Thank you for your insightful comments, Professor Harvey. I agree that the regulatory risk can be considered as one of the most crucial risks. Based on your lecture and previous literature review, I learned that the regulators are drawing more and more attention to the control of DeFi and blockchain. For example, China's central bank recently declared all transactions involving Bitcoin and other virtual currencies illegal. In addition, in the U.S., DeFi had officially entered the crosshairs of the US Securities and Exchange Commission (SEC). Moreover, major centralized and decentralized exchanges, such as Binance (the world’s largest cryptocurrency exchange), have recently been constrained to comply with KYC/AML requirements. These are all regulations or actions that contribute to an increase in regulatory risk.
Interviewee bio:
Haoxin(Peter) Yu is a Senior student at Duke Kunshan University, majoring in Data Science. He has strong enthusiasm for seeking interdisciplinary correlations, especially in the area of Data Science, Business, Fintech, and Psychology. He has also gained vibrant training in mathematics, statistics, computer science, and data analytics. During the study, he participated in multiple data-analysis-based projects, which enriched his experience in data mining, cleaning, and processing, as well as various quantitative models, machine learning. He had been doing research on traditional ESG Rating and is now doing research on risk exposure index on DeFi and blockchain. Besides, since he values the significance of integrating theories with practice and is willing to broaden his horizons, he has been applying the knowledge he learned to internships at consulting and technology companies.
4. Governance Policy on DeFi
William Zhao (Student):
The central bank of China recently declared that all cryptocurrency transactions are illegal, while the SEC in the U.S. is considering broadening the definition of “security” to include cryptocurrency in its regulation. These moves coincide with the regulatory risk you mentioned in your book. If you were a policymaker, how would you regulate cryptocurrencies?
Prof. Campbell Harvey:
This is something from William about the SEC broadening the definition of security, and I kind of highlighted that I would talk about this one. They went after Coinbase because Coinbase was going to offer an account where you can deposit a US-dollar-backed token and earn 4%. In decentralized finance, it's routine to get more than 4%. So, Coinbase probably gets eight, and then they would pay four, and then they will profit from the other four. So, anybody can do this today to get a much higher rate of return than what you would get at a traditional bank. At the bank, maybe you get 1% on your savings, maybe, but more likely less than 1%. Why are the rates so low? Well, the bank needs to pay for a lot of things like their building, their employees, and they need to make a profit. In these platforms, it's just an algorithm, so they can afford to pay more and I think that the US went after Coinbase because they could go after a successful firm that is centralized. It’s much more difficult to go after an algorithm. How do you do that?
So there will be this regulatory risk, and within DeFi there will be some disruptions. The stable coins are primary targets for this, and it'll be interesting to see how it plays out. But, it is important that it is played out because right now it’s just too much uncertainty. We need some sort of regulatory guidance right now. In the US, the securities act of 1933 doesn't mention cryptocurrency. So, this needs to be a work-through, and we'll see what happens.
William Zhao (Interviewee comments and/or follow-up questions):
Thanks for your comments on regulation! It is indeed important that we have some more certainty in regulation. A more recent announcement by Jeremy Powell saying that the U.S. will not ban cryptocurrency is very inspiring for the market.
Interviewee bio:
William Zhao is a junior at Duke University majoring in economics and math and Director of Research at Sciecon CIC. His academic interests include development economics, cryptocurrency, and fintech. He works with Prof. Luyao Zhang, Prof. Yulin Liu, Prof. Fan Zhang, and Prof. Kartik Nayak on the transaction fee mechanism design of Ethereum. William intends to pursue a Ph.D. in economics after his graduation.
5. Software developing for DeFi
Zesen Zhuang (student):
DeFi is booming today, with a wide range of dApps deployed on blockchain supporting its ecology and development. Given the current blockchain infrastructure, what do you think will be the bottleneck that will limit the next step in the development of DeFi applications?
Prof. Campbell Harvey:
DApps are decentralized applications. Instead of linking with a centralized company (i.e. WeChat), DApps are completely decentralized and they communicate via algorithms with others. The bottleneck in terms of developing DApps right now is scalability. Nevertheless, people are working on these applications right now since they know that the bottleneck of scalability will be transitory and go toward a much faster system in 2022.
Thank you very much for your forward-thinking response! I strongly agree with your point about the current bottleneck of DApps being scalability. From my observation, most of the current DApps are lightweight, single-function apps, rather than heavyweight, feature-rich apps that exist mostly in the centralized space due to scalability problems on decentralized platforms (Ethereum for example). Besides, do you think the bottleneck of scalability is primarily limited by current theories, such as consensus mechanisms, protocols, etc., or by the computer hardware?
Interviewee bio:
Zesen Zhuang is a Data Science student of Duke Kunshan University. He has a solid foundation in Computer Science and excels in Data Science courses. His areas of interest include reinforcement learning, decentralized applications, and game theory. He is working under the guidance of Prof. Luyao Zhang on the combined application of mechanism design and reinforcement learning. He is also involved in several projects at SciEcon CIC, for which he provides core technical support. He is the Director of Technology Development at SciEcon CIC, working on exploring the possibilities of decentralized networks.
6. Cryptocurrency as legal tender
Ray Zhu (student):
El Salvador became the first country to accept Bitcoin as legal tender about two weeks ago. Also on the same day, BTC crashed to its lowest in a month. This reminds me of the volatility risk, regulatory risk, privacy debate, and all other negative sides of a cryptocurrency you mentioned in your Coursera course. People from El Salvador also complained about the high cost of converting BTC into USD, which is high at 10%. What do you think about adopting cryptocurrency as a legal tender?
Prof. Campbell Harvey:
El Salvador has already pegged currencies to the US dollar, so it was kind of unusual to go to an alternative legal tender, which is Bitcoin. They could do it because the president is very popular and can run this experiment. Transactions in bitcoin are very expensive. So, during my course there was a transaction: somebody moved $5.4 billion of bitcoin within a half hour and the fee was $19, which is just a trivial fee for a $5.4 billion transaction. That $19 is fine for 5.4 billion. But, for a cup of coffee, that doesn't work. So, what El Salvador is doing is to use layer-2 technology. I described this earlier where essentially a secure wallet is set up, you can do the transactions within the wallet at little or no cost, so I think it's an interesting experiment. But to hold the bitcoins is much more volatile than holding the El Salvador currency which is linked to the dollar. If El Salvador gets in a distress, they could break the peg and many countries have had that problem where you thought your currency was pegged to the dollar, all of a sudden, there's a revaluation and you are Haircut. But I think it's an interesting experiment, and we'll see how it plays out.
Ray Zhu (Interviewee comments and/or follow-up questions):
Thank you for your answer! I also think what happens to El Salvador is very interesting and may lead to more exposure of both the advantages and disadvantages of cryptocurrencies. But I am still doubting that Bitcoin as a legal tender could work well. Seeing its volatility (it is even much more volatile than the US stock market), how could it be a stable unit of count, medium of exchange, and store of value?
Interviewee bio:
Ray Zhu is an economics student from the inaugural class of Duke Kunshan University. His research interest lies in financial investment, cryptocurrency, blockchain, and behavioral economics. Living in a community of over 70 countries and regions, he is also trained as a global leader with interdisciplinary and intercultural skills. Ray is the President of SciEcon AMA, leading a young scholar team to interview distinguished professors and entrepreneurs from the top universities worldwide. Being a pioneer in undergraduate research, He already has co-authored papers with Prof. Luyao Zhang submitted. He also serves in the student ambassador team at Duke Kunshan to spread China’s voice to the world.
7. Quantum Computing for DeFi
Yufan Zhang (student):
In the AMA interview published on SciEcon-AMA, you mentioned two courses you teach at Duke. One of them is Tech-driven Transformation and Business, where topics related to blockchains like quantum computing and brain-machine interface (BMI) are introduced. What’s your take on the potential impact that quantum computing can play on the DeFi and some other fields based on blockchain?
Prof. Campbell Harvey:
Quantum has nothing to do with the proof of work (PoW). If you think that a quantum computer can do proof of work faster, you are not understanding quantum computing. What quantum computing can do is potentially, in the future, recover your private key from your public key. The way that elliptic curve encryption works is that a private key just generates a random number. Thus, it is you that generates the number and then you pass it through an algorithm that is a one-way algorithm that delivers a public key and then a public address. You can share your public key and address, then somebody will send you Ethereum or Bitcoin, which you can then control with your private key. Therefore, quantum is a risk that, eventually, it will be able to go from public key to private key. However, in my opinion, that’s not a really big risk because we’ll know well in advance when the quantum computers get close. What you are going to do is to do a transaction and send your theory or send your bitcoin to yourself. When you do that, you will sign the transaction with a signature that is quantum-resistant or quantum-proof. These signatures exist today.
Therefore, it is not a risk for all of those that have their private keys, but a risk for all the keys that have been lost. If you lose your key, it’s lost. A huge number, which could be 5 million, of Bitcoin out there are lost. The keys are gone. Suppose that there is a firm doing quantum computing, which is a massive opportunity where you go in and you are able to recover these Bitcoins. It’s not illegal while people might say, that was my IP address. However, they cannot prove it. Therefore, I think that’s a big possibility.
Thank you for your excellent comments about quantum computing and DeFi. I initially thought for industries based on cryptography like DeFi, quantum computing would be a potential threat. The idea you presented about the possible solution for us to using a quantum-resistant signature is very inspiring and interesting. I have found that there are a lot of platforms and companies providing quantum-resistant services. Research on quantum-resistant algorithms has emerged, such as lattice-based systems (NTRU), error-correcting code-based systems (McEliece), and multivariable-based systems. Also, I think once quantum computing really arrives, switching to a new encryption algorithm will be the consensus of all Bitcoin stakeholders, so it is logical to upgrade through forks. Moreover, I am very inspired by the idea that we can recover those lost Bitcoins.
Interviewee bio:
Yufan Zhang is a Junior student at Duke Kunshan University, majoring in Data Science. He has great enthusiasm for disciplines ranging from Data Science, Computer Science, and Finance. Yufan has a solid knowledge foundation of machine learning, graphic design, and programming. He has participated in several interdisciplinary research including Summer Research in Computer Vision at the Institute of Software, the Chinese Academy of Sciences, and the GTML group under the advisor, Prof Huansheng Cao. Yufan is a powerful force in the workplace and research teams using his positive attitude and tireless energy to encourage others to work hard and succeed.
8. Consensus Algorithm for DeFi
Xinyu Tian (student)
Consensus building is one important part of blockchain protocols. And there has been a lot of research under the presence of synchrony and asynchrony, such as the Byzantine Fault Tolerance blockchains. So I’m wondering how consensus mechanism matters for DeFi, and if there are any risks regarding the consensus building on blockchain and other decentralized ledger systems?
Prof. Campbell Harvey:
This is a great question. In my Coursera, I deal with a large list of different consensus protocols. This is a good area for research, because, in my opinion, the particular application that you are focusing on might have an optimal consensus mechanism. There might be one application that it doesn’t make any sense to do Proof-of-Work for, and another application where you’ve got delegated Byzantine Fault Tolerance. So there are many possibilities here and there are many different mechanisms.
I mentioned Proof-of-Stake and there’s another mechanism that is quite attractive which is called delegated Proof-of-Work. This is where you could delegate some stake to and it’s somewhat analogous to a mining pool. Even though you are small, you can delegate some stake and earn a reward based upon that. So one problem with Proof-of-Stake is that the rich get richer because they are earning all the fees. The more you put in, in terms of stake, the higher the probability that you are actually selected to mine a block. So the delegate Proof-of-Work got the advantage that anybody could actually get into this and start earning rewards.
So, I don’t have THE answer to your question, other than it’s likely that this is specific to the application. There are many, for example, centralized firms that use blockchain technology which have nothing to do with Proof-of-Stake or Proof-of-Work. They are using Byzantine Fault Tolerance and the sort of mechanisms in a private implementation of the blockchain, which makes sense for them. They are trusted already, doing business, and essentially in control of this. So you need to focus on the particular application.
And this is a very new technology so that’s why there will be many up-and-downs. And for some of the problems, they haven’t been solved yet. And consensus building is one of those problems.
Thank you so much, Prof. Campbell Harvey! I am currently working on research about consensus building on blockchains and I’m trying to apply interdisciplinary game theory to evaluate miners’ behaviors. And your answer has given me another view on the current blockchain protocols that we should not only analyze the consensus itself but combine the mechanisms with particular applications to research. In addition, I believe we will explore more on the consensus-building problems as you mentioned that there are many to solve and we are seeking the optimum.
Interviewee bio:
Xinyu Tian is a junior student majoring in Data Science at Duke Kunshan University. She is a full admission scholarship recipient and a researcher at the Environmental Research Center. She is passionate about the interdisciplinary study of Machine Learning and Intelligent Economics, as well as the disciplinary study of deep learning. And she is currently working on game theory problems and consensus mechanism design in blockchains, supervised by Prof. Luyao Zhang. Besides, she is also the Director of Communication at SciEcon CIC, working on SciEcon publications and assisting the AMA program.
9. Stablecoins
Saad Lahrichi (student):
What would happen to the DeFi ecosystem should a major stablecoin like Tether crash?
Prof. Campbell Harvey:
Let’s first take a step back to understand how a stablecoin would crash. USDC (Circle) for example, is guaranteed by Coinbase, who are keeping the dollars in very safe assets. As such, it would take for Coinbase to fail for USDC to crash. Another example is if we buy treasury bills and deposit the money in a bank. That is risky, since the bank could fail - break the buck in US slang, i.e, falling below $1.
We must keep in mind that USDC and USDT (Tether) are both coins of centralized companies. Although USDC has applied to become a bank, I personally think that they would fail because they are trying to become a “narrow bank”. For example, when a customer deposits money in a usual bank, the bank sends 10% to the Central Bank as a required reserve and keeps the other 90% for lending. In a narrow bank, all of the customer’s deposit is sent to the Central Bank as there is no borrowing. The savings rate would then be whatever rate the Central Bank would pay. This effectively and shockingly quickly translates USDC into a Central Bank digital currency. All they do is take the customer’s money and “park it” in the Federal Reserve in the US.
On the other hand, decentralized stablecoins, like MakerDAO, or Fei are completely decentralized, they have an established governance system, no centralized Dollar deposits (they are instead backed by other cryptos), and are over collateralized, making it hard to go after them.
I think that the stablecoin space is very interesting. US banks are even making their own stablecoins now. JPMorgan has actually been the leader in stablecoins for major banks as they realized that it is a great idea. They want to use stablecoins internally to reduce costs for their customers and increase their profits. It is also the case, when I talk about stablecoins, that I talk about the free banking era in the US, where the banks in the US - not the Federal Reserve - issued the currency and it would be collateralized with gold, or silver, or equities, or US Treasury bonds. I think stablecoins are no different than these banks that issued the currency themselves.
The stablecoins I like the best are the decentralized ones like DAI and Fei. Fei actually was founded by a Duke graduate, so there is a good reason to like it. Again, this space is young and obviously in the regulatory cross-errors.
Thank you very much, Professor! Indeed, this is a very young space and it is still open to many innovations not only technical but also and especially regulatory. Recent reports have shown that the US is working on establishing a policy framework for stablecoins. When this happens, and stable coins are eventually treated as banks, it would be a major step towards the adoption of crypto and would attract more people to the space. While there is little reason for all USDT owners for example to withdraw their assets, I wonder what would happen if everyone decides to withdraw and Tether does not have all that money ($ 86 bn).
Interviewee bio:
Saad Lahrichi is a data science major student at Duke Kunshan University. He is passionate about Machine Learning and its interdisciplinary applications. Last summer, he worked as a member of Professor Ruth Day’s research team. The project’s aim was to study the psychology behind knowledge structures in learning. He has also joined Professor Luyao Zhang’s summer research team that explored applications of Machine Learning in FinTech and automated trading. He is now doing more research under Professor Zhang’s guidance on DeFi. Saad is currently serving as Assistant Director of Research at SciEcon and is also a member of Professor Mustafa Misir’s research team on Algorithm selection.
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC-BY-NC-SA 4.0)
