Associated Incidents
It was supposed to be a road trip, two weeks in a rental car through the USA. But when Sven Drewert wants to increase his credit card limit for the holidays, he gets a surprise. The bank declined the increase. The reason: His rating at the credit agency Schufa is too bad. Even though he always paid his bills. What kind of institution is this that uses its ratings, called scores, to pass judgment on almost every consumer? It's not a government agency, as some people think. Schufa Holding AG is a private company. Banks, mobile phone companies and online retailers who rely on reliable information about potential customers for their business rely on their ratings. The Schufa is not a monopolist: Other credit agencies such as Crif Bürgel, Creditreform Boniversum or Arvato Infoscore collect data from people and companies and assess their creditworthiness. But no other credit agency in Germany is as influential as Schufa when it comes to rating consumers. Participation in social life is hardly imaginable without credit bureau information: no mobile phone contract, no rented apartment, no bank loan, no paying on account in the online shop. According to the Schufa, it has information on more than 67 million consumers. This includes negative data such as payment defaults or entries from the list of debtors. But also so-called "positive information" about credit cards, checking accounts or real estate loans. A lot of this data is transmitted by banks and companies. The Schufa only saved positive characteristics of Sven Drewert, and still: A "satisfactory to increased risk" was the verdict, that the Schufa sent to the bank. Although credit institutions do not decide with whom to do business solely on the basis of the Schufa score, it can often enough lead to problems here. Drewert was lucky, his girlfriend was able to help out with her credit card for the US vacation. But to this day he doesn't know why the Schufa attested him an increased risk. Because how exactly Schufa arrives at its rating is a trade secret. You could also call it opaque. The algorithm is like a black box: it is only very vaguely possible to understand how it works – until now. Journalists from SPIEGEL and Bayerischer Rundfunk were able to evaluate the Schufa data of more than 2000 consumers. We will not reveal a secret Schufa formula: The data set represents only a small, quite distorted section of consumers in Germany - men are over-represented, older people are under-represented. Nevertheless, the data provide an unprecedented insight into the heart of Germany's most important credit agency and allow important conclusions to be drawn about how the Schufa scoring works. They show that Schufa knows far less about many people than one might think - and yet dares to calculate an exact score. The data also provide indications of how age and gender may affect risk assessment. And what disadvantages consumers can have if companies use old Score versions.
Chapter 1: Positive features
still no contract The Schufa itself states: It only saved positive contract information from more than 90 percent of all consumers, ie the existence of accounts and contracts. "And that's a good thing," says the Schufa website, because that could make life easier for consumers. For people like Sven Drewert, that must sound like mockery, because positive features do not mean a perfect score. On the contrary : Of those consumers in the dataset for whom there are only positive characteristics, the Schufa certifies every eighth an "increased" or even "high" risk in the Schufa base score. Whether this judgment is based on the number of their current accounts, for example, or perhaps the second credit card was the decisive factor, consumers do not find out from Schufa. It is also noticeable that Schufa knows surprisingly little about many people. Schufa has stored a maximum of three items of information from economic life for almost a quarter of the people in the data set - for example the establishment of a current account and the conclusion of a credit card and mobile phone contract.Otherwise, she only has general data such as addresses, alt he and gender. And although the Schufa knows so little in some cases, it assigns ratings that can become a problem for consumers. A 20-year-old employee who made his self-assessment available to the OpenSchufa project found out what that can mean. The Schufa certifies that he has an "increased to high risk". In addition to his age, gender and the number of addresses, their judgment is based solely on the existence of three current accounts, one of which has existed for several years. The man is not alone in this. In the data set there are more than 20 consumers who Schufa certifies as having an "increased risk", although their financial history contains no more than three entries - all positive ones. The Schufa also doesn't know much more about Sven Drewert, who was denied a higher credit card limit because of this judgement: In his self-disclosure three current accounts are listed, one of them with a credit card. That's enough for the Schufa to declare it a risk. According to data protection officials, it happens again and again that credit agencies such as Schufa are missing important information from a person's credit history. And it happens that companies request a credit report on people that Schufa doesn't think they know - although the people have long been stored in the database. This happened to one of the journalists involved in this research a few years ago: After moving, a provider refused him a landline connection because the Schufa allegedly did not know him and subsequently returned a very pessimistic score. Only later did it turn out that the data had not been assigned correctly due to the move. The Schufa itself only noticed this when asked. (Read the whole story here.) In our There are now similar cases in the data set: A man from southern Germany wants to sign a mobile phone contract. The provider asks the Schufa about their payment behavior. The Schufa initially says they do not know the man and does not transmit a score. She obviously takes the request as an opportunity to create a profile for the man. In his credit history she enters: "Request for a credit check before concluding a telecommunications contract". When the mobile phone provider asks again shortly afterwards, the Schufa now knows the man and plays back a score. The value cannot be based on much more than that The man's personal data and the fact that an inquiry was made about him - the score is miserable, the contract does not come about. The man has had a checking account for a long time, which the Schufa should have known. The account is now also appearing in his Self-disclosure. The chronological sequence documented on it suggests that he himself pointed this out to the Schufa. Ironic as that may sound: if consumers discover gaps in their data extract, they should inform the Schufa about it. This applies even more urgently to negative information, should they can be wrong: If a company reports payment disruptions, this negative feature attracts even the highest score ch below. The Schufa saves such entries for several years - even if the person concerned considers the matter to be over for a long time. There are several people in our dataset who, through fault or no fault of their own, have defaulted and fought their way out: they went through consumer bankruptcy and used a statutory portion of their income for up to six years to pay off the accumulated debt . As a reward for this so-called period of good conduct, a judge released her from the remaining debt. This should enable people to restart economically – actually. At the Schufa, the remaining debt exemption is stored for another three years. Over time, the data suggest, the impact of this information gradually weakens. But especially at the beginning, a re-entry into social life is hardly to be thought of: With a negative entry in the information for landlords, you will hardly find a new apartment in a big city. And the number of mobile phone providers willing to do business should also be manageable.
Chapter 2: Young and male
so unreliable It is actually wrong to speak of the Schufa score. Schufa calculates different values for different sectors such as banks, telecommunications companies or mail order companies. And over the years, new versions of these industry scores have been added, in which Schufa has adjusted its formulas. At the beginning of 2018, three different versions were in use. In addition, companies can have their own individual Schufa score calculated. The basic functionality of the scores is similar. From the combination of their stored data, the Schufa calculates a value between 0 and 10,000 points for a person. How exactly which feature is included in the calculation is secret, the weighting differs from industry to industry and from version to version. One thing is clear: the higher the value, the better. In the information provided by a young man from the data set, Schufa calculates a score of around 9493 points for the banking industry. That doesn't sound so bad, but the Schufa concludes that his probability of fulfillment, i.e. the probability with which he makes payments in accordance with the contract, is 93.79 percent. In words, this means a "satisfactory to increased risk". If the man had achieved just two points more, the percentage would be 95.62 percent and he would only have a "low to manageable risk". Because what many do not know: For each score variant, only a limited number of around 15 different fulfillment probabilities are transmitted. In this case, about 93.79 percent or 95.62 percent. Ratings in between, about a probability of 94 percent, do not exist. As a rule, the consumer does not find out anything about all this. The classification of the Schufa is uncompromising here, it draws hard limits and does not indicate any statistical error ranges. Little things can lead to a consumer slipping into the next category down, getting a more expensive loan or having to opt for a prepaid mobile phone because the provider doesn't want to give him a contract. But not only information about checking accounts and mobile phone contracts can play a role. Especially in older score versions, characteristics that cannot be influenced at all also affect the Schufa rating: A 25-year-old man must expect different treatment than a 65-year-old woman. This can be seen from the so-called "risk in the data types", which is shown in the Schufa reports. Under "General data", Schufa assesses the risk of consumers based on information such as date of birth, gender and the number of addresses stored. If you have a minus sign in your information, the Schufa certifies that you have an "above-average risk" in this category. In the data set, for example, younger people are more often rated poorly than older people, and that with otherwise similar characteristics. It is also noticeable that in this category Men are more likely than women to have minus signs. This is particularly common among young men in the data set. And the data also suggest that moving a lot also has a negative impact on the rating. Nevertheless, even with the more than From the 2000 pieces of information available to us, it is not possible to say conclusively how exactly the factors affect the respective score and whether Schufa includes them individually in the calculation or combines them.By the way, it is not forbidden for age and gender to be included in the score at all - as long as there are actually statistically verifiable differences.Schufa emphasizes only such Use information permitted by law to calculate credit scores. The General Equal Treatment Act, which is intended to protect consumers from discrimination based on age and gender, does not apply to credit bureaus.
Chapter 3: Contract or not
it depends on the version After all, information on age and gender is no longer included in certain industry scores. The latest generation of Schufa scores, "Version 3.0", was introduced in 2017 and for some sectors only takes into account information from a consumer's economic life. The new version offers more precise predictions, Schufa writes on its website. But the data analysis shows: many Apparently, companies are not yet using the supposedly more accurate forecasts.In the available data, the majority of companies requested old score versions in 2018. They can still call up version 2 scores, the formulas of which have now been in use for ten years, until June of this year version 1 was even used, which Schufa had introduced in 2001. This is not a disadvantage for consumers per se.The data analysis shows that some consumers have a better rating in the old score versions than in the new version 3. It works but also the other way around: a car sharing provider, for example, always asked for an old score from the retail sector for its customers. In the most extreme case, the company was informed that the consumer represented a "very high risk", although the latest version of the score only sees a "low to manageable risk". There are also these differences in the bank score. However, consumers cannot see which version of the score a bank is actually using before concluding a transaction. The extent to which the calculation methods differ can be seen when you compare the different versions of the bank scores for all individuals in the dataset: In the following graphic, each line connects the fulfillment probabilities of one and the same person – on the left the bank score according to version 2, on the right in the current version 3. Schufa and its customers always point out that the score it transmits is only one of many decision-making criteria when it comes to granting a loan, for example. Companies also have the option of directly using the positive and negative characteristics stored by Schufa. And often a bank has known the person for a long time anyway and knows, for example, how much salary goes into their account each month. However, bank insiders who wish to remain anonymous confirm that a Schufa score that is too low can lead to the bank refusing a loan. In the case of US traveler Sven Drewert, the bank also referred to his Schufa score when denying him the higher credit card limit. We presented the findings of our evaluation to Schufa and asked them for an opinion. We asked about the different score versions and how age and gender affected the score. We wanted to know why even people without negative characteristics get bad ratings and what uncertainty is associated with scores that are based on little information. The Schufa replied in a nine-page letter, but did not want us to quote from it or reproduce the content analogously.
Chapter 4: Better scoring is possible!
Consumer and data protection advocates have been demanding more transparency from Schufa and other credit bureaus for years. "Scoring providers should disclose to consumers the essential characteristics on the basis of which they are scored and, if possible, their weighting, in an understandable and comprehensible way," says the recently published report of the Advisory Council for Consumer Questions, an advisory body of the Federal Ministry of Justice. However, business secrets should not be violated by this. Gerd Gigerenzer is co-author of the report. Credit bureaus fulfill an important function, says the renowned scientist. But he also says: "It is time that we put the interests of consumers before the business interests of credit bureaus ." He would like "consumers to have the right to understand how their value comes about." Consumers do not yet have this right. Schufa always responded to calls for more transparency in the same way: with resistance. In the meantime, consumers are finding out even less than earlier if you order a free self-assessment: When the EU General Data Protection Regulation came into force in May 2018, Schufa obviously took the new rules as an opportunity to cut back the free self-assessment to the bare essentials An example from the USA shows what a better explained scoring could look like. The credit agency Fico there explains to consumers in a brochure how the so-called Fico Score works. "How you paid your bills is a very important factor of the Fico Scores," it says there, for example. The item "Payment history" therefore accounts for 35 percent of the Fico score. And the consumer learns that it is also received how long ago the missed payment was and how large the amount due was. For comparison: the Schufa informs consumers that "Information about undisputed, due and repeatedly reminded or titled claims" are processed. But the weight with which this information is included in which score variant remains unclear. Clear graphics? Everyday language? Also missing.