[ad_1]
Schufa is the most influential credit agency of Germany. It collects data from more than 67 million consumers and forecasts its payment history. Banks, mobile service providers, online retailers and homeowners incorporate Schufa's badessments into their decisions.
According to privacy advocates, data stored on a person is often incorrect or incomplete. Both can have a serious impact on the so-called score of a person – and thus on a consumer's decision to contract a telephone contract or on the terms of the granting of a license. a loan.
How exactly does the Schufa notation is a trade secret. In 2014, this was even confirmed by the Federal Court of Justice. Yet consumers want to understand how properties and behaviors affect their judgment. During this research, it was possible for the first time to take a complete look at the Schufa database and thus to better understand how it works.
OpenSchufa Project
This was made possible by the OpenSchufa crowdsourcing project. In February 2018, the Open Knowledge Foundation Germany (OKF) and AlgorithmWatch organizations invited consumers to request their own free self-badessment from Schufa and make it available to the project. (Read here how to request your own information for free.)
Participants scanned their information or photographed it. On the initiative's website, they had the opportunity to black out personal information, download documents and provide more personal information in a questionnaire. Some 2800 people took part in the campaign and provided their data.
OKF and AlgorithmWatch transferred the information via image format text recognition into a machine-readable structure. The data thus prepared were made available to SPIEGEL journalists and Bavarian radio. Similarly, journalists had access to the pseudonymized information of the questionnaire participants, as well as to the original blackened documents, so that all the results of the evaluation could be verified.
The information does not show a representative sample of the Schufa database: the number of men involved in the project is significantly higher than that of women and the elderly are underrepresented.
Work with data
With 2800 unrepresentative Schufa information, no algorithm can be reconstructed yet. But we can draw a lot of conclusions while the Schufa notation works.
If you look at the Schufa evaluation component, for example, which includes only general data such as age and bad, you can see that
- that young men are much more likely to have higher than average risk than older men.
- In addition, it should be noted that Schufa has stored in many consumers only three pieces of information or less of their economic life.
- and that some consumers are subject to increased risk even without negative entries. (Read all the results of the evaluation here).
All considerations could not take into account the 2800 self-declarations. On the one hand, some pages were not scanned or photographed correctly, which resulted in text recognition errors. On the other hand, Schufa has reduced the scope of its free information during the project. It is obvious that the entry into force of the General Data Protection Regulation (DSGVO) in May 2018 has been an opportunity to reduce substantially the copying of data.
The old information under the Federal Data Protection Act (BDSG) contains all Schufa scores published so far, as well as the current badessment of solvency. The new DSGVO information since May 2018 contains only the evaluations actually sent to the applicant companies during the last twelve months. The current badessment of creditworthiness that they do not deliver. This should be known to consumers.
Disclosure: Schufa's competitor, Arvato, belongs to the Bertelsmann group. His daughter, the magazine publisher Gruner + Jahr, is involved in 25.5% stake in SPIEGEL-Verlag. In addition, the Bertelsmann Foundation is one of the financial backers of the AlgorithmWatch organization, which shares the OpenSchufa project. Whatever the case may be, the SPIEGEL report is published independently.
Source link
