Mexico has two Credit Information Companies (SIC): Credit bureau Y Credit Circle. They are entities regulated and supervised by the National Banking and Securities Commission (CNBV), and a large part of their value is to present financial entities and commercial companies with decision tools for granting credit.
This tool has been critical in containing credit risk, in addition to allowing an entity to establish a risk model with relative simplicity, in turn increasing the supply of credit and with the potential to improve credit conditions.
There are alternatives to the Credit Bureau, but not as reliable
It is true that there are emerging industries that seek to generate alternative data that can be used to determine a person’s credit risk, especially when said person has no credit history. However, in my experience as the founder of Prestadero, this alternative data is not very predictive in determining how likely it is that someone will pay off their loan. That is, if a person uses Android or iOS, or if the pixelation of their phone is higher or lower, it hardly has a correlation with the performance of a credit.
Similarly, if a person uses the keyboard to switch between fields instead of the mouse, the speed of typing the request, whether the browser is up-to-date, among other factors, have turned out to be smoke in terms of credit risk predictability. Do you know why? Because there is no better indicator of whether a person will pay their credits, than the indicator of whether they have been paid correctly before.
I think it should come as no surprise that this is the most predictive factor. Even when you include Artificial Intelligence strategies and Machine Learning In your risk model, the most predictive variables in terms of credit default risk remain the same as always: repayment of previous credits, ability to pay, level of indebtedness, etc.
How to improve the Credit Bureau
For this reason, the Credit Bureau has been and will continue to be fundamental in people’s credit risk analysis, however, it does not mean that it cannot improve, and here are some ideas:
1. Have credit grantors update in real time
You see, when credit grantors request a person’s history, we see an old snapshot of the person’s situation. It is likely that this snapshot does not contain the updated data and the real situation of the applicant. That is, I could punish the rate or the decision to approve a credit because I notice that the balance of a credit card is very high vs. your limit. But it turns out that this balance is not the current one, that the person is even “total”, and that he simply had bad luck that the card issuer reported on a date where that balance existed. If the issuer updated the information in real time, with each movement of the applicant’s account, the credit history would be a much more useful and reliable tool.
The example also applies the other way around. In other words, a grantor could authorize a loan from an applicant whose “snapshot” of his history is impeccable: perfect and on-time payments, healthy level of indebtedness, among others. But it turns out that said applicant has already stopped paying and has taken other credits that are not displayed even because the reporting period of said grantors is monthly. So, someone who knows how to take advantage of this weakness in the system could request many loans in a month with the intention of accumulating the largest amount of funds, and put the grantors in check during this entire period.
So, in summary, grantors could be forced to update the information via API when certain important changes occur in an account, such as: delivery of a new credit, account delinquency, or application of a payment.
2. That more data be included to validate identity
The grantors may be obliged to send more information to the SIC, such as email, electronic devices from which the credit is requested, geolocation and IP address, among others. This type of information could help in the battle against phishing. That is, if someone requests credit from a new device, it could generate a possible alert to further review the request.
In the same way, the SICs should be updated to send more information, even obtained from other sources to be able to complement the identity information, such as: reputation of the IP address, validation of the applicant in the IMSS/ISSSTE, etc., validation billing before the SAT, voter codes, passport numbers, income validation. These data must currently be obtained by grantors from other sources when we could obtain them from the SICs themselves, already regulated entities that could provide greater value (and even generate greater income for them).
3. Find smarter validation ways to check your history
The autograph firm is in decline. Although this is a mechanism for an applicant to validate himself before the grantor to consult her history, I think we should think about better tools. The “Authenticator” or PIN are better options (when you enter data from your own history to validate yourself, for example), but they are already outdated. Why not think about a two-factor authentication with control of the requester or some other more modern mechanism? For example, if when a grantor checks your history, he asks the applicant to his cell phone or registered mail for a unique use code, or even through an application of the SIC itself. This would strengthen validation, combat identity theft, and put more control of their finances to the consumer.
Of course, there are many more options for improvement, and some of the ones I mention do not even depend on the SICs themselves, but on the regulator (CNBV, Banxico). And if this article sounds like a complaint, it is not, because without the SICs, Prestadero could hardly exist. But I think that if we have an opportunity to improve, we should always look for it for the good of the country and the improvement in access to credit that all of us in the industry know must increase.