Chen Daofu: Theoretical Thinking On Digital Finance

On November 14, it was co-sponsored by the School of Finance of Renmin University of China, the China Fiscal and Financial Policy Research Center, and CITIC Securities Co., Ltd., and hosted by the International Monetary Institute (IMI) of Renmin University of China, the Financial Technology Research Institute of Renmin University of China, and the Research Department of CITIC Securities The "Financial Technology Innovation Forum and Monetary and Financial Roundtable·Special Seminar" co-sponsored was successfully held in Beijing. This closed-door seminar is themed “Technology Empowerment and Inclusive Frontiers”. Chen Daofu, deputy director of the Financial Research Institute of the Development Research Center of the State Council, attended the meeting and delivered a speech titled "Theoretical Reflections on Digital Finance." Data and technology have no value or power in themselves; the latter is given and revealed through use. Digital finance is actually the combination of a new generation of information technology and finance to form new formats, new businesses, and new models, which is a reshaping of finance. Financial supervision is to balance the rights, obligations and abilities of each subject in the process of rebalancing. Facing the technological reconstruction of finance and a more complex market-oriented life system, the risk management of regulatory authorities may need to be more modest, improve and change themselves.

Chen Daofu, deputy director of the Financial Research Institute of the Development Research Center of the State Council

The following is the full text of the speech:

The author thinks about some of the logic behind Internet finance and digital finance from a theoretical perspective.

What is digital finance and what does it bring us?

In recent years, many concepts have been constantly changing, expressing roughly the same meaning, but with slightly different focuses, such as Internet finance, financial technology and now digital finance. The meaning expressed by these concepts is to connect data, technology and finance through a new generation of information technology to form new financial services, organizations and models.

The current so-called new generation of information technology generally refers to technologies such as big data, cloud computing, blockchain, and artificial intelligence. Finance has been using technology and data since its inception. After combining with big data and new generation information technology, the first is to continue the original relationship. Existing financial institutions and businesses are re-articulating them using next-generation information technology. Fintech updates. The second is financial reshaping driven by big data and information technology. Both approaches are progressing, but the reinvention of finance has greater social impact and far-reaching consequences, accompanied by certain technological paradigm changes.

Finance essentially deals with relationships between people and uses trust to reallocate resources in time and space. It is a social technology, regardless of what kind of information technology is used. However, the development and introduction of information technology has lifted the original time and space constraints to a certain extent, and finance can realize its functions in new models, organizations and business methods. In fact, digital finance is not just a combination of new generation information technology (data collection, transmission and processing technology). Behind it are the development of finance itself such as financial engineering and financial behavior, as well as the research results of human behavior and artificial intelligence. Intelligence combines machine learning and integrates all existing theoretical achievements of mankind.

In short, behind digital finance is actually the process of deconstructing and reconstructing finance from elements, links, businesses, organizations, and industries. The original forms of products, enterprises, institutions, and markets have changed, the boundaries have become blurred, and new connections and combinations have been explored. . Digital finance is also a process of returning to people (customers) as the center. Customers can receive more considerate and efficient financial services. Finance can become ubiquitous and ubiquitous, hidden in various production and life scenarios. In other words, financial services are becoming more subjective and thoughtful, and also more objective and rational, mainly based on the evolution of data and rules. The development of digital finance has accelerated finance to become a living and complex system.

What is the source of digital financial value?

Data is the perspective through which we understand the world

Data is a lens through which we understand the world. From a data perspective, "the world is data." The real world in which humans live, including the physical world and human society, is the world reflected by data (converted into conceptual text, images, sounds and other data using certain mapping standards, and recorded digitally). This is also a world where humans use data to describe, identify, study, and hopefully transform objects. In this sense, it can be said that data is just a "medium", ultimately for humans to better understand, transform and utilize the real world. From a human perspective, the real world is the original source and final destination of data.

Specifically, data is a mapping of the real world (physical world and human society) in which humans live. It is described by "numbers" (data description methods, including numerical methods (numbers) and non-numerical methods, but numerical methods cannot ultimately be converted into numerical methods. Therefore, it can be considered that the world can ultimately be described using numerical methods. - Editor's Note) to describe real-world states, transitions, and relationships. In this way, the world is divided into two parallel worlds: the real world and the digital world. By mapping the real world into the digital world in some way, we can more accurately describe the real world and simulate its operation according to familiar cognitive methods. In order to better serve mankind, we can find possible laws and create a virtual digital world. Through low-cost simulation, we can simulate the best way to realize human intentions and achieve the final result in the most efficient and feasible way. real world. This method can bring together human wisdom on a wider scale and over a longer period of time to achieve real-world goals with higher efficiency (low cost). From the perspective of data collection, aggregation, analysis and application, data are always based on people's certain cognition and abstraction of the described "object" in a specific technical context. It can be considered that the real world is a holographic world, and the world displayed by data is a world extracted through observation from a certain perspective. The process of digitization is the process of selection. Choosing nature means certain standards and reflects certain value tendencies.

In fact, the real economy is in a cycle of "investment, production, distribution, consumption, and savings". Through monetization and financialization, the monetary and financial system maps the exchange of barter, savings, and investments to the monetary and financial system (a type of digital space) to achieve reasonable pairing and response with the real economy. After the introduction of the digital world, more real economic activities, such as consumption and production, not just investment and saving behaviors, are connected and mapped to the digital space in real time, allowing for more reliable recording, connection and analytical processing. This can broaden the processing space of the original monetary and financial system and improve connection and processing efficiency. Therefore, the original monetary and financial system can be considered as the original digital economy. The current development of the digital economy has only changed the means of connection, recording, storage and processing, but it has not changed the behavioral logic and rights and obligations embodied in the data.

We can use the following figure to roughly express the relationship between the real economy, finance and digital technology.

Added value generated by digital finance

Changes brought about by significant reductions in marginal costs

Whether it is financial means or digital space, by mapping the real physical world to virtual space (financial or digital space), processing costs are greatly reduced. Although a large amount of infrastructure is required to collect and transform data, once the infrastructure is built, the marginal cost tends to zero, which is the so-called "zero marginal cost." Of course, the substantial reduction in marginal costs has also brought about economies of scale and corresponding natural monopoly problems. .

The first is due to the reduction of transaction costs (fees), which triggers changes in transaction and organizational forms. Whether it is the market, bureaucratic organizations or financial means, they are all designed to overcome high transaction costs and achieve better division of labor and cooperation in human society. "Zero marginal cost" will not only change the effective organization of financial institutions, financial industries and financial businesses, but will even trigger a revolution in the organization and management of the entire society. Retaining the autonomy of partners as much as possible, more flexible and closer external cooperation will become important features of organizations at different levels.

Second, the cost of applying existing research and operational processing will be significantly reduced. Anything that humans already know can theoretically be achieved through technological means. Artificial intelligence technology is not only machine learning, but also a low-cost application of human wisdom achievements so far. This will improve the level of human rationality and improve the efficiency of many-to-many matching in the market. The original institutional innovations introduced by simplifying the matching process will no longer be needed. Standardization and customization can go hand in hand, improving customer satisfaction and providing supply efficiencies.

Reduce information asymmetry, overcome human subjective uncertainty, and directly face pure unknowns

The world has both predictable continuities and unpredictable emergencies. The uncertainties facing humanity can actually be divided into two categories. One is the uncertainty of the objective world. Because it hasn't happened yet, humans are completely unable to recognize it. This part of the risk cannot be eliminated. Another type of uncertainty is caused by information asymmetry and cognitive inconsistency, that is, subjective uncertainty. The more "deep" you go into the object you want to understand, the subjective uncertainty will be eliminated to a large extent, and the only thing left in the end is "choice". Multidimensional data and real-time dynamic representation can greatly reduce subjective uncertainty, allow humans to directly face the unknowable unknown, and eliminate subjective uncertainty caused by information asymmetry and insufficient cognition.

The combination of low marginal costs, high processing power and reduced information asymmetry allows finance to begin to face and deal with areas that were originally considered high cost and high risk. For example, so-called long-tail customers and closed-loop capital circles can especially use more dimensions of data and stronger low-cost data processing capabilities to find specific cycles in greater uncertainty and achieve closed-loop operations, so that production can form. The self-circulating process of consumption chain, industrial chain, industrial chain and trade chain.

In addition, in the digital age, if the understanding of consumers and producers is accurate enough and the divisions are detailed enough, complete discriminatory pricing can even be achieved, ultimately achieving a resource allocation effect similar to full competition.

Digital finance can be embedded in production and life in real time

Big data is different from raw statistics. It has the characteristics of dynamic and timely feedback and is embedded in daily production and life. Describing and interacting with people or institutions in real time involves human cognitive and behavioral processes to a certain extent, and reflective theory comes into play. Cognition changes behavior, thereby changing outcomes, ultimately causing economic or financial systems to rapidly transform into complex systems. On the one hand, financial functions can be hidden in the scene, improving the autonomy and convenience of financial applications. On the other hand, the risk management model of traditional acquaintance society can be implemented in modern stranger society.

Of course, the development of digital finance will also bring new problems. First, it may lead to people’s belief in big data and artificial intelligence. This involves not only whether handing over cognition and decision-making processes to machines (artificial intelligence) is conducive to human growth, but also whether the capabilities of big data and artificial intelligence are as powerful as everyone expects, and where. Boundaries are. Artificial intelligence has realized the scientific research results of so many years of human society in the digital space. Whether machine learning can replace human exploration, whether the speed can meet needs, and whether there are boundaries are all worthy of further observation.

The second is the self-locking problem. Once the data is formed, there is a certain cognition behind it, and the data will strengthen this cognition, thereby hindering other cognitions, or making it more difficult for other cognitions to arise and deepen. This process becomes self-reinforcing and thus self-locking. Today, targeted push based on customer preferences increases stickiness and targeted advertising, but also limits the boundaries of human behavior.

Digital finance evolves based on data and rules, but there are multiple equilibria in the world, and evolution may also lock itself in. Whether the artificial intelligence-based mechanism in the digital space can achieve self-correction, break through possible self-loops and self-lock is the key to determining the robustness of this ecosystem.

Digital financial regulation

If we distinguish between the recording and implementation of data and technology, and the behaviors and relationships realized by humans using data and technology, then finance is the human behaviors and relationships generated by the application of data and technology in specific scenarios, especially trust relationships. among them. The development of financial technology has changed the means and accessible space of finance, but it has not changed the financial relationships that need to be regulated and adjusted—the rights and obligations of different participating subjects and “universal trust.” Due to the existence of varying degrees of "general trust", finance is different from other fields in the adjustment of rights and obligations.

It can be seen that data and technology should be regarded as the underlying support. What financial supervision needs to adjust is only the human behavior and relationships caused by the application of these data and technologies in scenarios. In other words, the essence and risks of finance are reflected in human behavior in scenario applications, rather than general data records and technical implementation. When artificial intelligence enables machines (technology combinations) to have human behavioral characteristics, similar to defining a company as a "legal person", artificial intelligence can also be regarded as some kind of abstract "person". The interface between digital finance and the real world is similar to the interaction between marketing and investment with consumers and investors in traditional finance, which is all conducted through various apps and websites. Behavior in digital finance is represented by various models and machine learning, and cooperation is represented by various interfaces (such as APIs). Traditional behavioral supervision also needs to be transformed into specification requirements for models, machine learning, and various interfaces. Therefore, in the development of financial technology, the focus of financial supervision is still the rights and obligations of "people" (natural persons, legal persons and "artificial intelligence persons") in scenario applications, rather than the specific means of data. Documentation and technical implementation. Of course, in addition to adapting to changes in form, you also need to adapt to changes in implementation form, and the above management also needs to be completed through big data and programming.

The widespread application of data and financial technology has improved financial efficiency and reached boundaries, but it has also brought about some new changes in risk characteristics. Data and technology have huge potential to unlock the value and certainty behind existing behaviors. Therefore, when the service group expands or the environment changes, financial risk changes are no longer linear, but will show huge non-linear characteristics. For example, digital finance serves a large number of tail customers, using data and models to find the low-risk portion of the original high-risk customer group. Under normal circumstances, in a reliable population, these data and models do identify low-risk groups so services can be provided at lower cost. But this group is a disadvantaged group after all and is relatively vulnerable. Therefore, as business interests continue to expand service targets, the risk characteristics of these customers will undergo fission. When the economic environment changes unexpectedly, the risk characteristics of this group also change non-linearly. So, are these risks absorbed by institutions through capital, or are they borne by institutions and passed on to the public through market decentralization? From the perspective of risk absorption by the whole society, the more people bear it, the total amount of risks that the system can bear will increase, but weaknesses may become shortcomings. The process therefore requires a better balance, which requires institutions to have sufficient capital to absorb extreme risks, while also allowing the financial system to develop more effective risk-sharing mechanisms.

Digital finance has obvious economies of scale due to its low marginal cost. Therefore, how to balance the imbalance of market power that may be caused by natural monopoly becomes prominent. In addition, digital finance is highly dependent on the collection of data. How to protect the rights and obligations of data collection objects, especially natural persons and legal persons, and how to deal with privacy issues have also become new issues. Finally, digital finance is highly dependent on information technology. In addition to paying attention to the relationship between people, we also need to pay close attention to the security and standard issues of data and technology to improve the security and convenience of docking.

Edited by Guo Ruihua

标签： #Finance #Digital #Cognition #Applications #Technology