Question 1: New technologies such as artificial intelligence, big data and blockchain have been mentioned repeatedly in recent years, but when it comes to specific applications, there are only a limited number of points that may come to mind, for example, blockchain technology is an advanced technology, but only Bitcoin can be applied at present. How should other areas apply decentralised blockchain trust creation mechanisms? Is there any other application of the new technology and data in other areas as well? For example, in the field of taxation and finance.
Gong Qiang: In general, the mechanism of blockchain has been available to us before, but it was limited by the lack of computing power at that time. We also know that the development of computers is growing exponentially, but when the computing power was not enough, trying to implement the application of consensus mechanism was not possible at that time. How can a computer with a 286, 386 or 486 put a consensus mechanism to use? Therefore, in that case, a series of problems arise. What is theoretically feasible is not clear to us whether it will actually work or not. In this case, why Bitcoin is remarkable is because it is a real practical application of blockchain technology, and after using it for a long time people feels that blockchain technology is reliable, therefore this is a reflection of its creativity. After the emergence of Bitcoin, although there were many different opinions about digital currency, it was also seen that such a technology could be grounded and could become a practical application. I personally feel that the application of blockchain technology is related to further algorithms, the underlying technology, but also to computing power and how it is calculated. In addition, the application of blockchain technology is now increasingly integrated with practice. In terms of blockchain, it is the theory that leads the technology. There is also the concern about whether people will be replaced with the advent of big data and artificial intelligence. By AI we mean human-written intelligence, which requires more labour, except that this labour is not the same as the original blue-collar labour, so it is still the intellectual input and drive behind the application of the technology. In this respect AI will really make a huge difference to the world.
In addition, the multi-party review and checksum of the blockchain is particularly important. It may not be evident in the government sector, but it is particularly evident in the supply chain. Normally, blockchain technology has information data on several aspects of products, funds, and logistics. Blockchain technology can be garbage-in and garbage-out, with inaccurate inputs and outputs that may not work. But if there are now multiple verifiable parties, suppliers, sellers, raw material suppliers and other data must all match up, and if they do not, it means that there must be something wrong with the data. What's more, because blockchain comparisons are traceable and verifiable, problems can be traced immediately. Blockchain technology can be used to facilitate multi-party verification, and secondly, it can be tagged. For example, once an invoice has been reimbursed, it is marked where it has already been reimbursed once and we immediately check for irregularities on the second reimbursement.
Zhang Bohui: I will probably address a few points, the first of which is about fintech. One of the ideas or research points in my previous research centre was about fintech, and I covered all these technologies in the system of fintech. The second point is that, for example, to launch a rocket, the bottom layer of the rocket needs fuel, and I personally feel that big data is to play the role of this rocket fuel. The internet, blockchain and cloud computing are all part of the rocket launch boosters. What is the engine of the rocket? It is such an algorithm and arithmetic power of artificial intelligence that acts as the engine of the entire rocket. All these components are very important for the rocket to be launched, but the fundamental building block is the data, without which it would be difficult to cook without rice, so we must combine the data.
Secondly, I have also been following academic research papers related to fintech or such technologies. Many research scholars focus on blockchain, and research papers involving blockchain-related research are more likely to discuss the price of bitcoin and ICOs and the valuation of the price of bitcoin, and then go on to explain why bitcoin has that price. Whether the price of bitcoin is artificially manipulated is one aspect. Another aspect is the price of the ICO, the success of the issue, and the elements that affect the issue. These two things are actually backed up by data - you can get data on ICOs, you can get data on Bitcoin.
There is also a type of research that specialises in the contractual mechanisms of blockchains. These studies are more of a theoretical model to determine whether such a mechanism design makes sense in capital markets or not. These are the economic, financial and accounting and financial aspects of the current phase of the study of fiscal and taxation aspects. Another aspect of research on blockchain is about machine learning. There are still quite a few application scenarios for machine learning, such as applications to predict the price of a stock. The overall finding of the study is that some articles can find predictiveness. However, most stability tests and heterogeneity analyses have actually found that using machine learning to predict future stock prices is actually not very robust, or the predictions are still inefficient. In addition to share price forecasts, there are also, for example, analysts' forecasts of future earnings, and there is some research in this area. As it is slightly less predictive, the overall requirements for machine learning are also relatively lower, and this is a study based on machine learning.
There are also very general studies, namely studies on big data. Big Data seems to be a new concept, but from the perspective of finance, accounting and economics, big data is probably something we have been focusing on for the past 20 years, only that it is not fully defined by the term. For example, micro-trading data in the stock exchange, for example, is so frequently traded and covers such a wide range of information that it can be considered big data in itself. Financial scholars or practitioners have been focusing on data mining and data discovery for the past one or two decades, looking for any possible data in the capital market that can contribute new information. Therefore, especially in the capital markets, you will find a lot of new types of data being mined to predict share prices and to predict company fundamentals. Behind this information is the big data technology, the algorithms, the same set of data, used by different people to achieve different predictions.
Finally, the last part. There are now many technologies that can be applied to government work, such as in taxation or overall macro market forecasting, to help us determine the direction of capital markets. There are many of these technologies, but first do we need to give it a very perfect requirement that it must be perfect to achieve what we want it to achieve? Technology is constantly iterative and changing, therefore it is possible that we do not need a technology to be particularly perfect. Whether the use of technology is useful, or whether it is beneficial or harmful, still depends on the subject who uses it. It is more important to look beyond the technology itself to see whether the subject of the use is a good or an evil one, since moral or intellectual discipline on the part of the individual or the company itself is actually more important.
Zhang Kezhong: I will share some of the theoretical, methodological and practical aspects of the new technological changes for us researchers. The first is that our mainstream economics has long been based on a fundamental assumption, since Adam Smith, of the role of the "invisible hand". This means that market mechanisms can lead to trickling-down effects, that is, to a trickle-down effect. Theoretically, technological progress and economic growth can benefit every group. However, the division of labour process leads to a large gap in income distribution because different groups have different degrees of technological mastery, especially if technological change and industrial transformation are relatively rapid. It is also an important phenomenon faced by developed Western countries and China.
When the gap in income distribution widens, technological advances can lead to groups of people becoming useless or idle. If they are idle, there may be hidden dangers and some changes will occur in the society. This is why we say that our income distribution centres are closely linked to the public finance system. Economic growth, technological progress is a problem that involves primary distribution and therefore has to be addressed by redistribution. Take, for example, our most fundamental policy: social security. What everyone is talking about now is Universal Basic Income, which means that everyone has to talk about basic security. At a theoretical level mainstream economics does have a great deal of explanatory power, but the process of technological progress is a hot issue that requires further explanation and response to mainstream economics in terms of, for example, the trickle-down effect, the marginal cost of zero-sum, the existence of incremental payoffs. At a theoretical level mainstream economics does have a great deal of explanatory power, but the process of technological progress is a hot issue that requires further explanation and response to mainstream economics in terms of, for example, the trickle-down effect, the zero-marginal cost, the existence of incremental payoffs, and more. Therefore, on a theoretical level, there is room for further improvement in both fiscal and financial matters, because economics is, after all, the study of human behaviour or behaviour between people.
The second point is at the level of approach. Technological advances have led to the deduction of our research from traditional normative analysis to empirical studies and then to randomised experiments. How are we going to use this experimental approach to explore cause and effect? Technological advances, the advent of big data, have now arrived at data-driven. It also means that our traditional process of empirical analysis considers causality and endogeneity, whereas we have to consider the issue of endogeneity. Whereas traditional research uses a sample approach, our research is now sample to total. The big data approach is driven by the fact that we can potentially find out what the causes are from correlation analysis, which is a further enrichment of the methods we use in the research process. In addition to the methodology, how to share data is a big challenge for us. How do we bridge the boundaries, including between government departments and enterprises? China still has a lot of work to do in this area.
The third is the policy level, particularly in the areas of finance and taxation. I will talk about the application of these technologies in the context of taxation. There are economic consequences that follow the application of technology. China is a developing country. China's Golden Tax III leverages the use of big data, cloud computing and even blockchain technology. In terms of the process from Tier 1, Tier 2 to Tier 3, the government's ability to collect taxes is quite strong since the full roll-out in 2016. Traditionally, developing countries have set nominal tax rates pretty high in the process of taxation because they are flexible in their policies. However, as technology advances and the ability to collect and manage tax increases, the actual tax burden becomes higher and the ability of companies themselves to avoid and evade tax becomes weaker in the process. Therefore, a higher effective tax burden has a significant impact on business investment and on the economy. How to optimise the tax system, this is something we need to consider. This means that advances in technology also present many challenges and considerations for our public policies.
For example, the digital tax. We have discussed this issue in our direct communication with the tax bureau, which is how the tax should be collected when the platform is in Hangzhou, but the sale or transaction is in other places such as Hubei, Beijing or Shanghai. This involves the principle of territoriality. The tax policy may then change significantly. In the context of global competition, there is a wave of tax competition between countries, that is, a wave of tax cuts between countries in order to attract capital and labour to their countries. As the capacity to collect and administer taxes continues to improve, the question of how to optimise tax systems and how to consider tax agreements and cooperation between countries rises to the level of Global Public Finance. In the global context, our fiscal cannot just reflect on the national and regional level. Especially in the context of China being a major country, it is a big question how China can formulate fiscal and tax policies to think about global governance in a big data-driven context. Fiscal, tax and some public policies do not just affect one country internally. With such a major country, with such a large volume of exports, our policies affect many countries. These changes have many theoretical and public policy implications that need to be explored.
Question 2: Professor Guo Feng has been doing research on fintech indexes, can you give us some context and application prospects?
Guo Feng: What we've just talked about is big data, or the commercial application of fintech. However, there is still a gap in the commercial application to be applied in the academic world. For example, there are about 100 people in this room, and most of them probably have only a laptop. Therefore, for such a basic device, it is very difficult to compare with commercial organisations and to do something with big data, which is very difficult. The real data is in the hands of these high-tech companies, or in government agencies, who have powerful computing power, and this is reflected in academic research, which creates a problem. When we talk about big data, when we talk about fintech, what exactly are we talking about? One set is to do some research work with big data, and another is to study what impact the application of big data has on the economy. For example, one approach is to use artificial intelligence and machine learning to study capital markets and to study credit problems; another is to study how the widespread use of artificial intelligence has substituted for employment, etc. These are actually two very different sets of approaches.
Why is this a problem? It is because many university researchers do not have or do not have the equipment or the skills to do so, so they can only study and discuss conceptual issues like big data on a superficial level. How to overcome this problem? It will require academia to work with technology companies and government agencies. Since I arrived at Peking University as a postdoctoral fellow, we have been working with Ant Financial as well as compiling something. At first, we compiled something called the Road to Internet Financial Development, and later on we adapted something called the Peking University Digital Inclusive Financial Index. We used hundreds of millions of individual micro-data from Ant Financial Services, such as the Alipay accounts used by tens of thousands of people in each region and the number of bank cards used within them, and dozens of other similar indicators to compile a measure of the extent of digital financial inclusive development in each region. The underlying data and personal microdata cannot be exported to the public because of the commercial confidentiality involved in this model of cooperation. But we can provide some ideas and together we can discuss the weighting and meaningful methodology of such an index, etc. This model overcomes the commercial confidentiality requirements that data cannot be leaked, including the increasingly stringent requirements of personal privacy protection in the future, while at the same time enabling technology companies to use the big data they hold for the development of the national economy. We feel that this is a very important approach.
Question 3: As the risk of local government debt is relatively high at present, can you talk about some ideas on local government debt risk identification, risk management and risk response from the perspective of big data?
Xu Wei: The government department is responsible for guiding the direction of the work, the academic department is mainly responsible for research, and the technology company is responsible for taking the technology further to the ground. Government debt is one of three major battles that are complex and difficult to tackle. You mentioned earlier that the risk is high, but this benchmark cannot be directly defined. Overall, the risk of government debt is still manageable. Only within manageable limits, there are a lot of problems in there and a lot of adjustments that need to be made. As a data provider or data processor, we have to find out how to screen the data, which are the risk points, which are controllable and which are risky issues, which is one of the core elements of our work. The second is that there are definitely several levels of debt risk, from the national level, to the provincial level, to the local level, and even to the county and township levels below, and the risks at each level are different, and we have to look at the risks at different levels when we do the data. Thirdly, because debt risk is a very complex issue, we cannot look at debt risk alone, as it is caused by a combination of factors. Some of these are due to the government's lack of financial strength, resulting in its higher risk rate; others may be due to excessive borrowing resulting in a higher risk. There can be a number of factors that create risk, so when analysing this, it is important to screen what exactly is causing the debt risk in a particular area. It is not the case that a large amount of debt makes the risk higher, but rather that it is important to consider a combination of factors, that is, exactly which factors should be taken into account, how they should be calculated, and how they should be calculated in order to have objectivity. This is something that we have been discussing with government departments and academics.
Question 4: The issue of income distribution is now very problematic and the income gap seems to be getting worse. Advances in technology may make the gap in income distribution increasingly wide, especially for those who do not master this technology, companies and other subjects. Can we use technological means and methods to solve the problem of income distribution, so that the problem of unfair income distribution can be solved by new technological means?
Wu Lei: My major is operations research. I have been pondering the issue of how models and data can be implemented in recent years, and I have recently done some consultancy reports on smart cities and the digital economy. I would like to talk about my thoughts based on this aspect. The first thing I consider is the feasibility and workability. What is hot right now is quantum computing, especially with the introduction of "Jiuzhang", what exactly is this thing? I often explain to students in my classes that, for example, if you break up a hundred numbers and then arrange them from largest to smallest, it would take about 700 or 800 steps to do it conventionally, but as a quantum computer it might only take ten steps. This may be a sub-management at the mathematical level, which is not quite as critical as the one that is already there now which we call a root type of addition. The biggest dimension of quantum computing is the addition of constants, which is the same as saying I keep it within ten steps, whether it's a thousand numbers or ten thousand numbers. Quantum computing is a very big breakthrough at the level of computing power, but why we seldom utilize it? For example, why bank cards only need a six-digit PIN. They give a very simple reasoning, which is to say that it is enough to guarantee your five-digit deposit with a six-digit PIN. In fact, it is a question of cost, which with quantum computers the cost would be high. In simple terms, under what conditions can we manipulate quanta? Absolute zero degree. If we need supercomputing, we need to have air conditioning, or like Google nestled it directly under Iceland, Alibaba placed its servers under Qiandao Lake, or deep in the mountains for physical cooling. There is a significant cost to the condition of absolute zero, which means there is a significant cost to bringing quantum computers into practical use.
Secondly, the cost of time and money. I had been in France for many years before, when I went to do a procedure for a permit to move and there were two dozen stamps on it. In a place as cutting edge as Europe, why would it take two dozen stamps and a total of three months to stamp the document? If someone wanted to forge this document, the cost of his time could take three months to get the thing forged. If we put in a high powered blockchain, artificial intelligence technology to complete this set of verifications it might only take a second. Regardless of the monetary cost of the investment, the time cost is reduced to just one second, and when this profit motive is large enough, it is natural that someone will undertake this, thus invariably expanding the risk when using technology for security control. There is no such thing as an absolutely secure technology, and cracks are bound to occur. When they feel there is huge profit in doing it, he will take the plunge. But if certain parts of the process, although manual, take a long time and have a low monetary cost, resulting in a final crime that brings only a very low profit. Artificial intelligence is only effective if it is on the ground.
Therefore, when using a new technology, we have to think more about being able to find a perfect entry point. We went to Hangzhou to research big data, and although Alibaba is doing very well, there are only about two dozen business scenarios implemented now, and they are still on a rotational basis. What they are always looking for is the entry point, that is, they must find a good entry point in order for the tool to be used on the ground.
Question 5: I would like to ask Professor Gong, you are a researcher in supply chain finance. Can the financial authorities make these financial funds more valuable by means of supply chain finance for the various subsidies and incentives given to micro market players' enterprises?
Gong Qiang: Let me expand on this a little. There are actually two things that are being discussed: one is what you mentioned earlier, and the other is the digital divide and the effect of the deterioration of income distribution due to digital technology. What is clearer to us is that, with the development of technology, people have found numerous ways to make the market more efficient. However, because different subjects are in a homogeneous situation, in general terms, technology is certainly beneficial for the market as a whole.
However, in reality, we know that there are different entities in the market, some of which have been able to make good use of the technology and some of which have not had access to it. We also know that it is now a time when the main body of the market is competitive, so this is a time when the strongest are always strong. We often say that in digital technology, only the strongest will win. Technology is advancing at a very low marginal cost. It is now difficult for small businesses to thrive and this has given rise to some thoughts now about the anti-monopoly of platforms. To put it another way, anti-monopoly is not necessarily about controlling monopolistic links, as we would like to think.
To give the simplest example, after Taobao and Jingdong came out, but now there is Pingduoduo. What else was there that we didn't expect? Tiktok and Kwai's live online sales now have an immediate impact on Taobao's monopoly on the digital marketplace, promoting digital market competition while also having a huge poverty alleviation effect. Many rural poverty alleviation channels are now selling through live streaming, bringing them a significant amount of sales revenue. At the same time, there are some people who are not benefiting from this due to the digital divide. This is a situation where we need to use technology to provide the basic security for these people with basic needs.
My personal view is that the initial stage of development is to enable the technology to survive, to where it is more profitable, and to get it settled first. Once the copyright is in place, the technology needs to be downward and finally applied to the detailed aspects of each specific area. On the financial side, for example, the downward movement of technology will lead to further improvements in income distribution. From the national perspective, some sectors need support. With digital technology, such support must be of a precise nature. In the past, the financial support was imprecise, such as the fraudulent use of grants that occurred in many subsidy policies, resulting in the inefficient use of funds in the end. The Chinese government, as a powerful regime, has played a major role in developing the economy and ensuring fairness. Therefore, we have great expectations for the development of the entire finance discipline.
Zhang Kezhong: The issue of targeting government subsidies and fiscal and taxation policies, that is, how to avoid waste, is a very popular issue in practice and in academia. In production network processes such as supply chains, which chain and point should the government subsidise to be the most efficient? It is also an ongoing academic concern: Why is there so much waste in so many subsidies at present?
Question 6: The use of big data in tax evasion is now also applied to social security management. We can see that the relevant departments have done a lot of work, but there seems to be a lack of a coordination mechanism between the various departments. Is there any national policy design, academic and practical thinking on the sharing mechanism of big data? The second question is about the last step of policy implementation, has the finance been implemented for the underprivileged personnel and micro and small enterprises? How does technology address the final step in practical applications?
Wu Tao: The problem of non-sharing of data between departments is indeed a problem in the use of our e-government. My personal feeling is that the use of big data in e-government applications is not depend on the technology. The experts are working on some cutting-edge technology, but we are using more mature technology. The difficulty lies in the application, and within the application, the difficulty lies in the sharing of data. Everyone wants to use the data of others but no one wants to copy their own data to others, making the coordination aspect particularly difficult. The central government is now addressing the issue of data sharing, including in the new management measures issued by the National Development and Reform Commission, which place great emphasis on this aspect. If there is a major project to be undertaken, it is important to set out in the project proposal how the data will be provided to others and what data will be required from others, as a necessary element to enable the project to be undertaken.
In the past, project proposals and feasibility studies were basically just a matter of writing about their own situation. However, it is still difficult to coordinate, and there are some special and objective factors involved. This is because some data is confidential, even if it may sometimes be confidential and sensitive among authorities. There is also the way in which the data is shared. As an example, instead of providing a data package, the public security department provides an API, which is also a solution.
Guo Feng: I would probably express some different views. For higher levels of management in government departments, they can ask subordinates to submit data. But does the government have the right to demand that enterprises hand over the data they have collected themselves? This is a question that requires legislation. Perhaps the government can require companies to provide such data on national security or whatever grounds, otherwise in most cases I don't think the government has the right to do so. There is a question involved, which is whether the data belongs to the individual or to the big data company? For example, if I buy a thing on Taobao, this shopping behavior and the information registered in my personal account, this information is my personal, but the value is generated only after the big data company through the collection and aggregation. If Taobao does not share the personal data of its users, then I don't think the government has the right to ask for this data. Unless not sharing this data would create a national security risk, the government could argue on this basis.
On the other hand, I do think it's important to think the other way around - why the requirement to share data? If academic data is only valuable when it goes together, but why is it necessary to force data from different institutions to be stitched together? I shop on Taobao, what gives it the right to share my data to another commercial organization? This is ridiculous. I shop on Taobao, and under the terms you can collect information about my transactions, but that doesn't mean I authorise you to share my data with another big data company or another financial institution at will. This boundary should be regulated by legislation as soon as possible. It's not just this issue.
Take an example that we have all experienced. You searched for a product on Taobao, whether you bought it or not, and after you left Taobao you went back to Baidu or Tiktok and they quickly suggested some items to you, asking if they were of interest to you, indicating that they had scanned your Taobao search history. And where is this boundary? There are two ways of doing this. One way is that after I have made a purchase on a particular website, the trace data from that browser is quickly passed on to another company, which I think is a relatively unlikely to happen.
Another possibility is that I have an application installed, which monitors anything I do on this phone, even my chat history and my browsing on other sites. This is illegal, but until now there has been no legislation against illegal data sharing practices, data collection practices. Therefore I think it is not how to promote data sharing, but how to regulate malicious sharing.
Gong Qiang: How can we protect privacy and data sharing? In theory, blockchain technology is currently somewhat viable. However, the problem with blockchain is that it is still not mature enough. I believe that later on, when the blockchain is mature, it can be used for information sharing. As long as there is an agreement that I can share the data , or that I can access the data without copying it. In this way we can combine data sharing and privacy protection at the same time. The viability of the technology still has to be seen when the technology is on the ground.
Question 7: The difference between data as a factor of production now and in the past is that the people and companies that generate data are also data users. Under the circumstances that industry data acts as its result, the introduction of DCEP at the national level and the preparatory process for the National Data Centre will give rise to many new business opportunities. If data is interpreted as a factor of production, its distribution becomes important. Banking and finance are in fact the chain of distribution. In the context of blockchain or big data, how can data as a factor of production be distributed more appropriately and efficiently?
Zhang Kezhong: My comprehension is that the market mechanism is at work. Public resources are shared, private resources are shared in a way that privacy is protected by legislation. On this basis, it is considered a market mechanism if some processing of the data has been done on a legal basis to benefit market demand. That is, the government's data goes to the government and the market's data goes to the market. Making data work based on price mechanisms cannot be separated from basic market laws. We can think of data as an element, which is scarce and needs to be processed. Because the information obtained by data companies is useful to users, it is up to market mechanisms to work. Government data, in my opinion, still needs to go public. In many developed and developing countries, government data is available for the public to study.
Wu Lei: I think a more reasonable direction is to attempt to establish data trading centres in various regions. This includes Guizhou, Guangdong and Hunan, which have all established such data trading centres. Land can be traded, financial products can be traded, and data can certainly be traded. Once the trading system is established, data sharing is equivalent to corporate mergers and acquisitions, as long as it is reasonable and legal, it can generate value and everyone is willing to do it. The two databases merge together, just as one business takes over another. This actually happened in Europe a few years ago, a girl who sold all her data, which was a big event at the time. There may be some individuals who feel that their data is important, but there are others who believe that his data can generate value. Many people are not allowed to donate their organs to others after their death, but their organs can be donated to an institution, as long as they have signed an agreement authorising the organ donation. A more reasonable mechanism to promote the use and development of data would be to establish a data trading system. I think the state could support and explore such a mechanism.