论文部分内容阅读
A large number of big-data credit reporting institutions are surfacing in China as consumer finance， online lending and other emerging sectors grow larger. However， China must address six key problems before it can establish a diversified and multi-layered credit reporting system.
First， the country’s regulators must ensure the quality and authoritativeness of the data. Compared with the wellestablished People’s Bank of China’s Credit Reference Center， big-data credit reporting institutions obtain a greater variety of data from a larger amount of sources. Nonetheless， the data collected by the latter may not be as reliable in terms of quality and authoritativeness as the former. Based on a survey of major big-data credit scoring companies， the U.S. National Consumer Law Center published in March 2014 a report titled Big Data， a Big Disappointment for Scoring Consumer Creditworthiness， saying that 50 percent of the data obtained by credit scoring companies had errors. The report criticized big data’s susceptibility to inaccurate information， comparing its use to the “garbage in， garbage out” maxim.
Another factor is different credit ratings for the same person. The accuracy of predictions by big-data models depends on two factors： data and algorithms. Different big-data credit scoring companies adopt different algorithms and collect data from different sources. Among the current eight pilot credit scoring firms in China， Pengyuan Credit Service Co. Ltd.， China Chengxin International Credit Rating Co. Ltd. and Intellicredit Inc. use traditional algorithms and mainly collect financial and public data. Meanwhile， Zhima Credit， Tencent Credit Services， Shenzhen Qianhai Credit Service Co. Ltd.， Kaola Credit Service Co. Ltd. and Sinoway Credit all use data that they’ve accumulated from their respective businesses. As a result， different credit scoring companies may provide contrasting reports for the same person.
Third， privacy protection and information safety should be secured. According to the Regulation on the Administration of Credit Investigation Industry， the collection and application of personal credit information must be subject to the approval of those who are rated for credit. When making reports and inquiries to the People’s Bank of China for personal credit information， commercial banks must abide by this regulation， which clearly defines what information should be reported and which situations allow access to such information. However， bigdata credit reporting relies on a large amount of online transaction records and information from social networks. These sources of information have no clear boundaries for the protection of privacy and therefore the risk of privacy breaches is greatly amplified， making it difficult for citizens to protect their rights. Fourth， the availability of public information and the exchange of information throughout various organizations should be ensured. With the above analysis， most of the current personal credit scoring institutions in China collect a large amount of data from their respective businesses， meaning that these companies hold a monopoly over those sources of information. However， big-data credit reporting requires the sharing of information in order to work as intended. At present， no clear law， supervision or technology standards have been set on whether information can be exchanged among different institutions， what information obtained must be subject to authorization， and how to ensure that the information will not be misused during and after exchanges. Furthermore， it is not yet established that administrative information on commerce and industry， taxation and the administration of justice can be obtained in a sustainable way. Credit reporting institutions now collect such information from varied sources， so the costs of data collection are high， and the quality and sustainability of such data cannot be guaranteed.
Fifth， problems involving social security and unfair competition caused by the abuse of information should be resolved. Among the eight pilot personal credit reporting institutions in China， Zhima Credit， Tencent Credit Services and Kaola Credit Service have been in fierce competition， introducing their own rating products and applying these products in sectors such as finance， shopping， job hunting， car renting， apartment renting and hotel check-ins. However， it is unclear whether the reports produced by these institutions can truthfully represent one’s personal credit. Furthermore， the reliability of the technology used to obtain such information still needs to be tested. Unrestricted commercial applications are likely to increase security risks and consumer discrimination.
Finally， credit scoring institutions should maintain their independence. In a narrow sense， credit scoring around the world refers to activities providing information to lenders in order to control risks. Credibility of this activity is maintained only when those institutions are independent third parties following the principle that those who collect information have no relation to those who produce it. However， most of the credit scoring institutions in China are not independent third parties： Their data come from parent companies， while their sister companies are engaged in online lending. Their scoring results are highly correlated to their respective businesses and risk judgments， but the correlation of such scoring results to other businesses needs to be verified.
To realize the goal of establishing a diversified credit reporting system， China must improve its credit management laws and regulations， build a unified system for credit data， speed up the construction of a platform for various public credit information and strengthen the supervision of the credit reporting industry through concrete measures.
First， the country’s regulators must ensure the quality and authoritativeness of the data. Compared with the wellestablished People’s Bank of China’s Credit Reference Center， big-data credit reporting institutions obtain a greater variety of data from a larger amount of sources. Nonetheless， the data collected by the latter may not be as reliable in terms of quality and authoritativeness as the former. Based on a survey of major big-data credit scoring companies， the U.S. National Consumer Law Center published in March 2014 a report titled Big Data， a Big Disappointment for Scoring Consumer Creditworthiness， saying that 50 percent of the data obtained by credit scoring companies had errors. The report criticized big data’s susceptibility to inaccurate information， comparing its use to the “garbage in， garbage out” maxim.
Another factor is different credit ratings for the same person. The accuracy of predictions by big-data models depends on two factors： data and algorithms. Different big-data credit scoring companies adopt different algorithms and collect data from different sources. Among the current eight pilot credit scoring firms in China， Pengyuan Credit Service Co. Ltd.， China Chengxin International Credit Rating Co. Ltd. and Intellicredit Inc. use traditional algorithms and mainly collect financial and public data. Meanwhile， Zhima Credit， Tencent Credit Services， Shenzhen Qianhai Credit Service Co. Ltd.， Kaola Credit Service Co. Ltd. and Sinoway Credit all use data that they’ve accumulated from their respective businesses. As a result， different credit scoring companies may provide contrasting reports for the same person.
Third， privacy protection and information safety should be secured. According to the Regulation on the Administration of Credit Investigation Industry， the collection and application of personal credit information must be subject to the approval of those who are rated for credit. When making reports and inquiries to the People’s Bank of China for personal credit information， commercial banks must abide by this regulation， which clearly defines what information should be reported and which situations allow access to such information. However， bigdata credit reporting relies on a large amount of online transaction records and information from social networks. These sources of information have no clear boundaries for the protection of privacy and therefore the risk of privacy breaches is greatly amplified， making it difficult for citizens to protect their rights. Fourth， the availability of public information and the exchange of information throughout various organizations should be ensured. With the above analysis， most of the current personal credit scoring institutions in China collect a large amount of data from their respective businesses， meaning that these companies hold a monopoly over those sources of information. However， big-data credit reporting requires the sharing of information in order to work as intended. At present， no clear law， supervision or technology standards have been set on whether information can be exchanged among different institutions， what information obtained must be subject to authorization， and how to ensure that the information will not be misused during and after exchanges. Furthermore， it is not yet established that administrative information on commerce and industry， taxation and the administration of justice can be obtained in a sustainable way. Credit reporting institutions now collect such information from varied sources， so the costs of data collection are high， and the quality and sustainability of such data cannot be guaranteed.
Fifth， problems involving social security and unfair competition caused by the abuse of information should be resolved. Among the eight pilot personal credit reporting institutions in China， Zhima Credit， Tencent Credit Services and Kaola Credit Service have been in fierce competition， introducing their own rating products and applying these products in sectors such as finance， shopping， job hunting， car renting， apartment renting and hotel check-ins. However， it is unclear whether the reports produced by these institutions can truthfully represent one’s personal credit. Furthermore， the reliability of the technology used to obtain such information still needs to be tested. Unrestricted commercial applications are likely to increase security risks and consumer discrimination.
Finally， credit scoring institutions should maintain their independence. In a narrow sense， credit scoring around the world refers to activities providing information to lenders in order to control risks. Credibility of this activity is maintained only when those institutions are independent third parties following the principle that those who collect information have no relation to those who produce it. However， most of the credit scoring institutions in China are not independent third parties： Their data come from parent companies， while their sister companies are engaged in online lending. Their scoring results are highly correlated to their respective businesses and risk judgments， but the correlation of such scoring results to other businesses needs to be verified.
To realize the goal of establishing a diversified credit reporting system， China must improve its credit management laws and regulations， build a unified system for credit data， speed up the construction of a platform for various public credit information and strengthen the supervision of the credit reporting industry through concrete measures.