An excellent Account Information Service is based on the accuracy of the categorization of transactions

In 2021, Creditinfo Estonia received permission from the Financial Supervision Authority to start offering account information services in Estonia, which later expanded to the markets of Latvia and Lithuania. Today, we have been offering the account information service on the market for almost two years. The account information service is based on the PSD2 directive. We have access to the transaction data of customers of banks and financial institutions using a secure data transmission channel and customer consent.

Account information categorization is the first and most trivial account data processing that creates customer value. In addition to the primary value, categorization is also an input for all subsequent, significantly more value-creating services (for example, debt risk assessment). Without categorization, each time finding, analyzing and displaying value from account information becomes too resource-intensive, so the end user would have to wait a relatively long time to get a result from their data.

Unfortunately, categorization is worthless if the accuracy and quality of the categories are low. Of course, every transaction on a bank account is not an input for assessing a person’s credit risk. When determining credit risk, it is critically important that the accuracy of the categorization of transactions required for analysis is as high as possible. This is to prevent credit losses for companies and overdue debts for private individuals, directly affecting both interest groups’ reputations.

The main input from categorization is related to income

 

The primary input from the account information for credit risk assessment is salary/income and the volume of financial obligations (loans, installment payments, leases, etc.) per month. In addition, various red and green indicators affect a person’s credit risk. For example, casino visits and bailiff payments can be classified under red and insurance charges under green.
To ensure the accuracy of the categorization, Creditinfo has given the first priority to categorizing transactions important for credit risk assessment across the Baltics. However, today, we can state that the overall accuracy of categorizing the account information service offered by Creditinfo across the Baltics exceeds 90%.
 
A more accurate percentage value can only be estimated by looking at the categorization of a specific bank account since the accuracy of the categorization is directly related to the transactions that the bank account reflects.
 
Accurate categorization of account information is also essential for ensuring know-your-customer (KYC) and anti-money laundering (AML) rules for all companies to which KYC and AML rules apply to a greater or lesser extent. For example, too much cash mobility in an account can mean potential money laundering. There is not, and should not be, a definite rule as to what amount constitutes money laundering in the case of a large amount of cash in the account. Many companies operate in a sector where a lot of cash moves. However, this does not make these entrepreneurs suspects of money laundering. If the cash movement is justified, then the doubt is also grounded.
 
In summary, it can be said that the bank statement is a valuable new data collection that helps to assess a person’s credit risk better. The basis for a more accurate evaluation is categorizing bank account transactions of excellent quality. At the same time, it must be remembered that achieving 100% categorization accuracy is impossible. Service providers are constantly changing; people go on trips, new companies are born, older companies disappear, purchases are made in various domestic and foreign online stores, etc. These are all reasons why there are always companies whose payment transaction categories cannot be specified as soon as possible.

Visit creditinfo.ee/en for more information.

Ivo Vallau

Open Banking Product Manager, Ceditinfo Estonia.