Data transactions, data rights and data factor market development

Tian Jietang, Deputy Director and Researcher of the Innovation and Development Department, Development Research Center of the State Council

Data is a new production factor and the new "oil" that drives the development of the digital economy. General Secretary Xi Jinping has repeatedly stressed the need to "build a digital economy with data as the key factor." Clarifying transaction rules, improving standards and measures such as data ownership, open sharing, and transaction circulation, and accelerating the cultivation and development of the data factor market will help give full play to the role of data as a basic resource and innovation engine, and is also an important part of the market-oriented reform of factors under the socialist market economy system.

Practical Exploration and Results of Data Trading in my country

1. Local governments actively explore data trading

Data trading is a basic way to promote the circulation of data elements in the market under market economy conditions. With the rapid development of big data technology and applications, various parts of my country have carried out exploration and practice on data trading in various forms. In April 2015, the first big data exchange in China, Guiyang Big Data Exchange, was approved for establishment. In the following years, Wuhan, Harbin, Jiangsu, Xi'an, Guangzhou, Qingdao, Shanghai, Zhejiang, Shenyang, Anhui, Chengdu and other places have established big data exchanges or trading centers to provide data trading services. At present, there are more than 20 data trading institutions in my country, all of which are coordinated by local governments or national information centers, and a number of data operation service companies such as AsiaInfo Data, Jiujifang Big Data, Shuhai Technology, and Zhongrun Puda provide technical and operational support.

As the pioneers of big data trading, these data trading institutions have not only carried out useful explorations in practice for data trading where the rules are not yet clear, but also tried to formulate relevant rules for data trading and put them into practice, accumulating experience and lessons, and also achieved initial results. Taking the Guiyang Big Data Exchange as an example, it has successively formulated a series of trading rules such as the "Interim Management Measures for Data Confirmation of Rights", "Data Trading Settlement System", "Data Source Management Measures", "Data Trading Qualification Review Measures", "Data Trading Specifications", and "Data Application Management Measures". Although there are still many controversies in specific practices, this spirit of courage to explore is still worthy of full recognition.

2. Comparison of the two data trading models and their effectiveness

From the practice of data trading institutions in various places, there are currently two main trading models, which are also the two mainstream ideas for developing data trading institutions. The first is the data matching trading model. This model is a bit like a traditional commodity market, so it is also called a "data market". Under this trading model, data trading institutions mainly trade roughly processed raw data, without any pre-processing or in-depth information mining and analysis of the data, and only sell it directly after collecting and integrating data resources. Many exchanges or trading centers use this trading model as the basic development idea in the early stages of development. The second is the data value-added service model. Data trading institutions do not simply match buyers and sellers, but according to different user needs, they clean, analyze, model, visualize and other operations around big data basic resources to form customized data products, and then provide them to the demand side. From the practical results of various places, most data trading institutions have chosen the trading model of providing data value-added services instead of direct trading of basic data resources after many explorations.

There are two main problems with the data matching transaction model: First, this type of matching transaction requires the acquisition of a large amount of data resources, and it is often difficult to achieve effective personal information protection. Some scholars have pointed out that there are a large number of gray and black transactions in the data trading market, which seriously affects the in-depth development of data transactions. Second, big data itself has the characteristics of heterogeneity and low value density, making it difficult for most data demanders and suppliers to reach a price consensus. For customers, a large amount of "roughly processed" data has little significance for business decisions or research. Whether it is government or corporate needs, accurate and effective data may only account for one million or even one billionth of the total data volume, and the subsequent extraction and analysis means a lot of time and processing costs.

Relatively speaking, the data value-added service model has two advantages: First, data value-added service agencies extract low-density and high-value data from big data on behalf of customers, saving them a lot of time and analysis costs. For most small and medium-sized enterprises, professional talents who can meet data needs are relatively scarce, and in-depth mining and analysis of raw data require additional talents or technical investment. Therefore, directly purchasing data products that have been processed by targeted means can save a lot of expenses and have a high cost-effectiveness. Second, service providers that provide data value-added services need to ensure the legality of the data, reducing the legal risks of data demanders. In the current situation where some laws and regulations are not yet complete, data demanders sign contracts or agreements with data value-added service providers, and the latter is responsible for ensuring the legality of data acquisition and processing, effectively avoiding problems such as data privacy protection that plague data transactions, and enabling the data trading market to operate effectively.

Difficulties in defining data rights and related disputes

1. Difficulties in defining data rights

Data is different from ordinary tangible materials. In the era of big data, the production characteristics of data have changed, making it difficult to define the ownership of data.

First, data rights are diverse, and different types of data have great differences in the content of rights. The subjects of data rights include natural persons, governments and enterprises. Personal data may contain personal privacy, and natural persons have the right to privacy over their own data. Therefore, the rights of natural persons to personal data are intended to protect their interests in self-determination of personal data, thereby preventing the infringement of personal personality rights and property rights due to the illegal collection and use of personal data. Government data is generally considered to be a public resource, and the public has the right to know, the right to access and the right to use. Commercial data includes the intellectual property rights, trade secrets and legal rights of market competition of enterprises. From the current data-related laws, personal data is a clear legal concept, and there are clear legal concepts and normative systems at home and abroad. In the context of the discussion of government data openness, government data is also an important object of rights. Relatively speaking, commercial data is relatively vague and has not yet become a strict legal concept.

Second, the data production chain includes multiple participants, and the rights and responsibilities need to be divided among the participants, which leads to difficulties in definition. Unlike other property, the entire life cycle of data is controlled by multiple participants (data providers, data collectors, data processors, etc.), and each participant assigns different values ​​to the data in their respective links. In most cases, the data control processor (such as the network platform) needs to collect, process, handle and analyze the data in order for the data to play a role and generate value. Therefore, the data provider's rights to the data require the support and cooperation of the data control processor to be effectively exercised. It is not feasible to grant a certain participant exclusive and exclusive ownership. It is necessary to negotiate and divide between participants such as data providers and data control processors to determine the boundaries and relationships between the rights. The content of data rights will also change with the changes in application scenarios, and even derive new rights content, making it difficult to agree on the ownership of rights in advance. In the era of massive data, the negotiation of the complex rights content contained in each data by the data control processor will bring huge transaction costs. A set of simple and feasible rules is needed to make data utilization possible.

Third, data is different from traditional common objects in all properties. At present, many people have proposed to establish ownership of data. The characteristics of data (such as infinity and compatibility) make the "ownership" within the extension of data property rights different from the definition of ownership in general civil law (the exclusive right to use, benefit from and dispose of property). For ownership in the usual sense, the owner has almost complete rights to possess and use the object, and generally there will be liability issues when the object is not properly kept and infringement occurs. Data rights have different controlling entities throughout the life cycle of data, and the right holder needs to bear more obligations and responsibilities. Not only must they be responsible for incidents such as data leakage and data infringement, but they also need to perform corresponding obligations in daily data collection and processing.