Personal Finance 201: Introduction to credit scores

OK, the credit score. This is perhaps the most controversial topic and most confusing element in the credit history universe. Humans in all stages of their history, have been easily fascinated by numbers, and today still retain a subconscious desire to summarize almost everything in hard numbers. What’s the temperature today? What grade did you get from the econ class? How many days are left till summer? Don’t we all love numbers!

It comes as no surprise then, that the quality of your credit profile could also be quantified. Statisticians, mathematicians, and credit analysts, using highly advanced software programs and statistics and probability theories, have been working together for generations to find a way to formulaically analyze even the most complicated credit profiles and produce a most simple output. The result of their research is a formula that generates a single number from each credit report at a given time, a number that represents the financial riskiness of the potential debtor. This is how the famed credit score came about.

Did you catch a huge lie from what I’ve written so far? It’s a pretty big one.

There is no such a thing as the credit score.

Wait, what?

Well, there is no such a thing as the credit score.

………………………………………………………………………………………………………………………………………….

It may be common knowledge in academia as well as in the business world that in order to draw a conclusion, you always have to make assumptions, and an assumption can affect the result to an extraordinary extent. The teams of mathematicians, statisticians, and credit analysts have their own assumptions, resulting in very different formulas to quantify one’s credit profile. As you can tell, this leads to a plethora of credit scores generated from each credit profile. Many of them are still being used by lenders nowadays. Each lender, upon receipt of a credit application, pulls a credit report from 1 of the 3 consumer reporting agencies, applies the credit scoring formula, and comes up with a credit score. Based on this credit score, the lender will determine if you’re too risky for them to extend credit to, or to offer a favorable interest rate to.

There is an issue with this approach. OK, say Sallie Mae uses their credit scoring model to determine that your credit score is 500, on a scale of 0-700. So how risky are you as a borrower? How cold or hot is 40 degrees, if we do not know the temperature scale?

In reality, Sallie Mae would have to rely on some sort of database of past borrowers to evaluate your credit score relative to others. If your credit score is better than 99% of their past lenders, that means you’re a low-risk borrower. So what if you are a new bank that offers credit cards but doesn’t have such a database due to obvious reasons? And what if the nature of credit products and credit reports change in the future as a result of innovations and regulations? Would you have to adjust your credit scoring formula? Do you want to have an in-house team of credit scoring formula developers just for these rare occasions? Or is it more efficient to leave this complicated mathematical task to the hands of external experts, like how Apple contracts Foxconn to produce their iPhones?

As it turns out, the outsourcing model is easier to maintain. Some companies have developed their own credit risk models and sold creditors licenses to use the models. Each model generates a different credit score as well as a risk level from a credit profile, depending on how the model weighs the factors in the profile. To make it more confusing, each company may have several models to evaluate risks for different purposes, such as mortgage lending and credit card granting, resulting in even more credit scores. Each lender generally has a contract with a company that owns a risk model. Each time there is an application for credit, the lender applies the selected risk model on the credit profile to generate a score. And since each lender has its own preferred risk model, you end up with a different credit score.

Your credit score depends on the scoring model and the credit bureau from which the credit report is pulled, and each lender has their own preference for these.

So really , there is no such a thing as the credit score.

While there exist so many different credit scoring models, fortunately, there are only a few that are most commonly used by lenders. Over time, there is one company that has proved itself as the most reliable credit risk model developer: Fair Isaac Corporation, or FICO. Even though FICO has multiple credit scoring models, they all have the same methodology with the exception of the weight attributed to each factor. An auto-enhanced FICO score would stress factors that have a higher predictive power for auto loan risk, and a credit card-enhanced FICO score would put more weight on factors that more accurately predict credit card risk. But the utilization ratio is calculated the same way for both these scores, for example. A useful thing from this consistency is that you’ll have a good idea of how to improve your credit score for auto loan purposes if you know how your credit card-enhanced score is affected by changes to your credit profile. FICO’s consumer division has a website called myFICO.

I have subscription to a MyFICO service that provides me with FICO scores from their mortgage-enhanced model. Last month, because of a new credit card account added to my credit profile, my score took a big plunge. If I were to apply for an auto loan, I know that my FICO score for the auto loan would decrease as well, though probably not as much. At the same time, my non-FICO credit score from another website showed a minimal decrease because it calculates credit utilization a different way from FICO. So I’d have no idea how much my auto loan FICO score would be affected based on this non-FICO score.

Because I am new to credit, I’m very curious about how my credit activities affect my creditworthiness, specifically my FICO scores, so I can develop my credit profile in such a way to help me get the best rates when I need an auto loan or a mortgage in the future. Whereas there are many types of credit scores out there, I only trust FICO scores. There are two rules you should know about FICO scores:

1. The range is from 300 to 850.
2. Unless there is specific mention of FICO, you can assume a credit score is non-FICO. All scores that are not based on FICO models are derogatively called FAKO’s.

I will continue talking about credit scores, especially FICO scores, in future posts. As usual, if you have questions on this topic right now, please leave a comment below, and I will try to respond as soon as I can.