First, the background of the birth of big data credit reporting
In the past two decades, with the reform of economic system, great changes have taken place in China's enterprise credit system, from large enterprises as the main loan group to small and medium-sized enterprises as the main loan force. In the face of new loan groups, banks and other financial institutions can't give enough funds, which leads to the increasingly serious "financing dilemma of small and medium-sized enterprises". As early as 200 1, Lin yifu published the article "development of small and medium-sized financial institutions and financing of small and medium-sized enterprises" in Economic Research, arguing that small and medium-sized financial institutions are more suitable for serving small enterprises, which laid a theoretical foundation for vigorously promoting the development of small and medium-sized financial institutions in China. The establishment of city commercial banks, rural credit cooperatives and small loan companies has led to a surge in loan products for small and micro enterprises in the whole society, greatly improving the supply of funds and greatly changing the financing environment of small and micro enterprises in recent years.
However, the emergence of small financial institutions has not fundamentally solved the dilemma of micro-financing. Compared with large enterprises, the proportion of funds occupied by small and micro enterprises is extremely low (about 30%), which is extremely disproportionate to their contribution to GDP (about 70%). In recent years, the survival pressure of small and micro enterprises has been increasing, the competition in traditional industries has been fierce, and the profit space has been squeezed. Credit sales make small and micro enterprises face cruel capital turnover pressure, and small and micro enterprises are closed down everywhere because of the broken capital chain. This situation has further aggravated the "reluctance to lend" behavior of financial institutions. The negative expectation of small and micro enterprise loans leads to the contraction of small and micro enterprise loans, and small and micro enterprises and small financial institutions are caught in a vicious circle. Small and micro enterprises and small financial institutions are at the bottom of the whole credit system.
The dilemma of small and micro enterprise credit seems to be very complicated, involving macro and micro behaviors, but in fact all the difficulties of small and micro credit are concentrated on one point: the funders think that they can't get loans because they can't see the risks of small and micro enterprises clearly, which is the so-called "information asymmetry risk"; Unable to identify the risk, the investor formulated a loan policy to avoid small and micro loans, which formed an "adverse selection", and small and micro loans stopped here and fell into the predicament of no money to lend. Both big banks and small financial institutions face the same problems, so there is nothing they can do about microfinance. This problem can be collectively referred to as the lack of social credit system leading to high credit risk.
The lack of credit system makes it difficult for investors to see the actual situation of small enterprises clearly, which has practical reasons. The internal management of small and micro enterprises in China is very arbitrary, and many transactions will not be recorded in a standardized way. Regular investors need to go through strict due diligence to investigate the first repayment source (repayment by business) and the second repayment source (collateral) and make a decision. This process can be called "credit investigation" or "credit review". As we have analyzed before, due to the lack of effective credit investigation and audit methods available to small and micro enterprises, this process is not only time-consuming and costly for small and micro loan projects, but also difficult to find accurate, true and valuable information, which hinders the success rate of credit for small enterprises. Similarly, for those private equity institutions and private capital institutions, they are unable to conduct relevant investigations, and lending can only rely on feelings and other means, which is more risky.
It can be concluded that the bottleneck of small and micro enterprise loans is "lack of efficient, low-cost and high-precision basic credit information service". It is conceivable that if the investor has the ability to accurately identify whether the small enterprise is credible in a low-cost way, and then take auxiliary risk control measures (guarantee, mortgage, etc. ), small and micro businesses will become profitable, capital channels will be opened, and small and micro credit will become smooth and orderly. Big data credit is produced under such a social background.
Second, explore several ways of big data credit technology
With the in-depth application of big data technology in various industries, using big data to open up credit channels has gradually become the mainstream consciousness of society. Credit service practitioners, government credit reporting agencies and internet finance companies have made unremitting explorations in this regard, hoping to find a quantitative in-depth evaluation method for small and micro enterprises. Let's analyze the characteristics of several main methods at present.
(A) quantitative credit evaluation (rating) model (from the inside out)
For many years, credit institutions, credit bureaus and rating agencies have been expecting to form a quantitative credit model, which can automatically generate rating results after importing all aspects of data into the model, prompting whether it is possible to lend. After long-term exploration, research and experiments, this ideal model has not been released. Some powerful investors in China have introduced the credit analysis models of famous consulting companies in Japan and the United States, but these models are not suitable for the actual situation in China and have not achieved the expected results-after importing relevant data, they can make reliable judgments on the repayment ability and willingness of enterprises.
Advanced foreign models and years of model exploration by domestic institutions have not formed a universal and effective quantitative model to judge small enterprises, mainly because the data quality of small enterprises in China is low. Because the enterprise data used at home and abroad are mainly financial statement data, and financial data are issued by accounting firms. There is a huge deficiency in China's credit system, and the audit report issued by accounting firms is almost made, with low credibility. For honest enterprises, this report has great reference significance, but for enterprises that deliberately defraud loans, there may be no flaws in the audit report. The exploration of various quantitative models has not achieved satisfactory results, precisely because the data they are based on are of low quality, so it is impossible to get really valuable information anyway. This method is basically declared invalid.
(b) External database access (from external to internal) mode
In the case of poor internal data quality, various institutions began to look for more extensive and reliable data sources, such as government data, tax system data, industrial and commercial information, business data of industry authorities, customs data, business data of various industry associations, and transaction data accumulated on e-commerce platforms (such as Taobao). According to these data, find the data related to an enterprise and make a comprehensive analysis. We call it "external-internal" data system, that is, enterprise credit service is no longer to extract data from the evaluated enterprise, but to realize it by using external data system.
The advantages of this model are: after the database system is formed, it will be very easy for a single enterprise to collect credit information, and the marginal cost of credit information service is extremely low and the speed is extremely fast. The direct benefit is that the charge of credit information service will be very low and the service volume will be very large. But this model also has its own disadvantages: docking multi-department data portal is a huge system engineering, and the cost of construction and running-in is very high. At present, except for industrial and commercial information, other departments' information is distributed in municipal departments, and the integration work is quite huge. In addition, the most serious problem is data quality. The business information submitted by domestic small enterprises is very random and compiled according to specific needs, such as tax avoidance, loans or other purposes. In order to encourage the development of local enterprises, some local governments give high tax incentives, such as approving a fixed tax amount, so that enterprises will not be required to declare truthfully. Therefore, the data collected from various departments may be far from the actual situation, and its credibility will be questioned if it is used as a credit information service. At the same time, not all data generated by enterprises are submitted to the outside world. In fact, the data submitted to the outside world only accounts for a small part, such as basic financial statements, taxes payable, etc., while most of the data that can explain the situation of the enterprise are deposited inside the enterprise, such as supply and marketing information, product categories, capital flow, etc., which cannot be found through external databases. Although the external database has a large amount of data, it is not enough for a single enterprise. If the e-commerce internal ecosystem data is relatively one-sided, since an enterprise will not only sell through one e-commerce channel, the single e-commerce transaction data is obviously not comprehensive enough.
If we weave the data network with external data, this network will be huge, covering almost all enterprises in the country. However, due to the lack of data about an enterprise, the data lines of this network are sparse, that is, the data grid is very large, and most valuable information about the enterprise is omitted, and there are too few effective information to draw credible conclusions. This is an exploration of establishing a credit data system from outside to inside.
Since the State Council was appointed by the National Development and Reform Commission to establish a national credit system, government credit offices at all levels have taken the lead in connecting all government departments with the data within their jurisdiction to form a unified credit information platform, which is operated by professional third-party companies or subsidiaries and publishes credit reports that meet social needs. In addition to the government, there are also social credit reporting agencies doing similar things, accessing some government data and operating. Judging from the current development, the most important information that this kind of credit information service can provide at present is industrial and commercial registration information and a small amount of information filed by various departments. This kind of credit service provides simple information and low fees, but it basically does not play much role in credit business.
(c) Single enterprise data credit service (from internal to external)
Another data credit service is to mine useful information from within the enterprise. From this point of view, this method is consistent with the traditional credit method, but the information collected and the analysis mode are different. Now some professional credit information companies are also developing data credit information methods from the inside out. The amount of data used in this method is not as large as that of social credit information (so it is called "small data"), but it is highly related to loans. On the basis of ensuring authenticity, a lot of valuable information ("big information") can be obtained, which are the most concerned information of investors in credit business. This service can help the funder to judge whether the enterprise meets the loan conditions in the shortest time, save a lot of research time and cost for the funder, and adapt to the efficiency requirements and risk control requirements of financing for small and micro enterprises.
Although the amount of basic data used in data credit investigation is not as large as that of government departments, all the information collected is the most relevant, which can capture the real business situation and repayment ability of enterprises. From the point of view of data network, the data network formed by this method is small (only suitable for a single enterprise), but the data "mesh" is just suitable for retaining a lot of valuable information about the enterprise, filtering out irrelevant information and interference information at the same time, forming a deep and high-quality credit report, which provides a reliable basis for credit decision.
The advantage of this data credit service is that it can be started quickly without long-term construction cost, and it is well adapted to the existing information base and social reality in China. The difficulty lies in how to gain the full trust of the borrowing enterprises and thus be willing to provide in-depth data.
Enterprise data credit technology is not a technical idea, but has begun a lot of practice. Data credit investigation has played an important role in guarantee business and small loan business, helping guarantee companies and small loan companies to investigate risks, clearly evaluate projects and improve business efficiency. I believe that with the change of market environment, more and more people will realize the value of this technology.