What exactly is a score? How do you find out your score? Can the scores be influenced? What's the deal with geoscoring?
These and other questions will be answered here.

What is a score value and what is it used for?

For common reasons, the ratings of credit rating agencies are familiar to a large part of the population: agencies such as Standard & Poor's rate the creditworthiness of sovereigns and large listed companies.
The USA has been rated ("AA+") for a long time. This means that the rating agency considers the probability of default on U.S. government bonds to be very low.

From the point of view of the holders, government bonds are nothing more than a loan to the government. The rating serves as a basis for decision-making for individual investors and entire markets: Is the current interest rate high enough in terms of the risk assumed? Is the risk for an investment even too great and should therefore refrain from buying government bonds of a certain country?
Banks in the mass lending business are also faced with these questions: In order to be able to make lending decisions, the most accurate possible information on the credit default risk of a potential borrower must be obtained.

At the same time, the practice of the credit business requires information that is as standardized and cost-effective as possible. Credit agencies provide this information to banks in the form of score values.
A score value can be represented as a percentage of 100, for example: If a consumer has a score of 100 percent, the credit risk is zero. The lower the score, the higher the risk.
Most banks set a threshold below which lending is ruled out. At the same time, different risk premiums can be set for score values within the acceptance criteria.

Score values are based on empirical data. This applies not only to credit scores, but also to score values in medicine or meteorology. The task in developing a score is to use mathematical and statistical methods to establish a meaningful relationship between the data available on a person and the probability of default.
In simplified terms, a credit score could be developed as follows: Of 500,000 people who have changed their checking account more than once in the past six months and had to service at least two current loans at the same time, 8 percent had a payment default in the twelve months following day X. In practice, comparison groups play a major role.

In practice, comparison groups play a major role: consumers are assigned to a larger group of consumers and their probability of default on the basis of their characteristics (number of current and previous loans, number of existing accounts, payment problems, etc.).
Credit agencies such as Equifax calculate score values on a daily basis and take new incoming data into account immediately.

Equifax claims to use logistic regression methods, also known as logit models, for its scoring models. For math buffs, these are regression analyses used to calculate the distribution of dependent variables.

Geo Scoring - How it works and how it is used

A correlation between certain characteristics of a person and the risk of default can be established not only on the basis of information on existing and previous contracts.
Data on a person's more or less private environment can also be used for this purpose. In the context of what is known as "geo-scoring," data on the residential environment in particular plays a role.
In terms of data protection and a number of other aspects, the procedure is not entirely unproblematic, but the statistical correlation cannot be dismissed out of hand: If you live in an area where a large number of your neighbors have payment difficulties, you are statistically more likely not to pay bills and the like than residents of a "good" area.

Credit agencies can evaluate various data as part of geo-scoring:

- Data from motor vehicle and household insurance companies
- Information on payment problems at an address or in a street
- Information on age structure, crime rate and level of education