FICO Scores: Uses and Misuses
Jacob Fuller and Eric Dawson
The Fair Isaac Corporation and FICO Scores
Debt is an important tool for both businesses and individuals. It can be used to leverage returns, raise capital or make purchases that would otherwise be unattainable by most. Before exchanging money for a promise to repay, lenders need to judge the creditworthiness of a borrower to accurately assess the risk involved and to determine the required rate of return on their investment. When the loan involves a corporation, ratings agencies such as Standard & Poor’s or Moody’s, will rank the credit risk using publicly available information. When the borrowers are individual consumers, another approach to measuring credit risk must be employed.
The Fair Isaac Corporation provides exactly this, a method for quick and accurate assessment of the credit risk of individuals. Other companies provide similar services, but the FICO score is, by far, the most popular indicator of credit risk for individual borrowers. Since consumers are not required by law to publicly announce income, expenses, cash flows or savings, it becomes much trickier to accurately predict an individual’s ability to meet debt obligations. The FICO score attempts to accomplish this by using complex statistical models based on information gathered from credit reporting agencies. While there are admitted drawbacks to the methods used in calculating these scores, a FICO score is seen by many lenders to be a critical piece of information when assessing credit risk of an individual.
This chapter is intended to gain insight as to how accurately a FICO score measures credit risk of individual borrowers. Additionally, it will address what affects any shortcomings of the score may have on borrowers, lenders and financial institutions as a whole. Do most people know what affects their credit score? If so, can a score be manipulated to reflect the most positive aspects of the borrower? While transparency in the marketplace is seen as an extremely important aspect of an efficient, competitive market, at what point does evaluating the risk of individual borrowers encroach on personal privacy?
Background on FICO Scores
Before the value of the FICO score can be analyzed, an understanding of how it is calculated needs to be attained. Since Fair Isaac’s exact models are proprietary and virtually impossible to access, only a rough outline of the scoring system is available. An approximate breakdown of the components used to calculate a FICO score includes: punctuality of payment in the past (35%), the amount of debt (30%), length of credit history (15%), types of credit used (10%), and recent searches for credit and the amount of recently obtained credit (10%) (Fair Isaac). Missing loan payments will clearly have a negative affect on a FICO score, but the ways in which some of the other factors affect scores are less obvious. For example, the total amount of debt a person has is not weighted as heavily as the percentage of debt used in proportion to available credit. Also, if two people have joint loans, any delinquencies will have a negative affect on both people’s FICO scores despite who was responsible. The nature of this scoring system creates an environment where the average borrower does not have a perfectly clear idea of how their activities and history affect their credit.
Nationally, FICO scores range from 300-850 and the distribution is skewed towards the higher end. Figure 1 shows this distribution. A higher FICO score indicates a lower credit risk. According to the Fair Isaac Corporation:
By law, credit scores may not consider your race, color, religion, national origin, sex and marital status, and whether you receive public assistance or exercise any consumer right under the federal Equal Credit Opportunity Act or the Fair Credit Reporting Act (myFiCO 4).
Some other factors not included in a FICO score are age, salary, occupation, employment history, child/family support obligations and rental agreements. While many see these restrictions as a positive in terms of equal opportunity lending, it could be argued that some of the restricted factors are useful in determining credit risk (my FiCO).
Figure 1. Source – The Fair Isaac Corporation.
Users of FICO scores
Almost all commercial lenders use some type of credit score before loaning money to an individual. There are many different credit scores, but the FICO score is the most widely used. When a borrower applies for credit at a bank, the lender will check the borrower’s credit report and FICO score. When the score is delivered, it may also include up to five ‘score reasons’ which will help determine why a particular score is a certain number (myFiCO). The FICO score is intended to give a “snap-shot” of the borrower’s current credit risk based on their credit history. Since the FICO score does not include some important information such as income, employment, or savings, the lender will typically request verification of this information.
Institutional Investors and Securitization
The process of securitization has grown exponentially in recent history. This process involves the purchasing, repackaging, and sale of loans to investors (usually institutional) in the form of bonds. A wide variety of individual loans are securitized, including credit cards, student loans and auto loans. The largest sectors of securitized loans made to individuals are mortgage and home equity loans.
If the securitized mortgage loans meet certain requirements, they may be guaranteed by the Government National Mortgage Association (GNMA, also known as Ginnie Mae). This means that if the bonds (backed by mortgage loans) default, the United States government will pay back the interest and principal to the investor leaving no credit risk. However, if the bonds are not backed by the U.S. government, the credit risk associated with the underlying loans remains. Since there are many loans that do not meet the criteria for federal guarantees, FICO scores become essential to the assessment of credit risk when investing in these types of securities.
The dollar value of financial assets secured by mortgage loans in the U.S. exceeds $6 trillion. This is more than half of the gross domestic product (GDP) of the United States. At the same time, “recent studies show that nearly 80 % of mortgage lenders use credit scores as the primary determinant of an individual’s credit risk” (Schrock 1). Since this enormous amount of debt has been issued based on FICO scores, it is clear that there may be major repercussions in the financial markets if these scores do not accurately measure credit risk.
The borrowers for which FICO scores are given are also users of these scores. While it is required by law in some states that borrowers be allowed to access their credit reports for free at a certain frequency, the actual credit scores are not included in this report. Borrowers may buy their scores from the various companies that calculate and track them. Individuals can use this information to gain a better understanding of their own credit risk. It may help them to decide whether or not they would be able to make certain purchases or apply for lines of credit. If borrowers know their scores, they also have a better idea of how they have been managing their credit.
Another important use of FICO and other credit scores is for individuals to check on the possibility that they have been a victim of identity theft. If someone has used another person’s identity to apply for loans, this information will appear in the credit report. For example, if someone uses a credit card under another’s name and does not make the payments, a delinquency will appear on the credit report and the FICO score will be lowered. If the thief does make payments, the score will not be much lower, but a new line of credit will appear on the report signaling an identity theft. With identity theft becoming more prevalent, the ability to check one’s credit report and score is an important tool.
Accuracy of FICO Scores in Determining Credit Risk
Do People Know What Determines Their FICO Score?
The first step in analyzing the usefulness of FICO scores is to gain insight as to whether or not people understand what a FICO score is, how it is calculated, or how their actions affect their FICO score. If people do not understand how or why their credit scores change, they may not act in the same manner as they would if they had more knowledge on the subject.
At first glance, maintaining good credit may seem simple, but merely paying all bills on time does not ensure a high FICO score. As addressed previously, there are more factors that contribute to the credit score than just payment history. Out of 118 people surveyed, only 58% knew what a FICO score was. Furthermore, only 48% of those surveyed claimed to have any understanding as to how a FICO score is calculated. 57% knew that payment history has a greater affect on a FICO score than the length of credit history, but only 42% knew that debt as a percentage of credit limits has a greater affect than length of credit history. The survey results suggest that there may be a significant deficiency of FICO score knowledge, especially among new users of credit, since college students represented the majority of the sample.
Inefficient lending may be taking place because there does not seem to be a connection between what people think they know about FICO scores and how they are actually calculated. Borrowers may be acting in a way that hurts their credit score without knowing it. If these borrowers had a better idea of how their score was affected by their activities, they would likely be more careful with their credit. Of people surveyed, roughly 19% carry balances on their credit cards even if they have the money to pay them off (for various reasons). These people’s credit scores are likely lower than if they did not carry balances because their outstanding debt is higher as a percentage of available credit. If they knew that a high debt to available credit ratio negatively impacted their FICO score, they might be more likely to pay off their debt with available cash.
Statistical Accuracy: FICO Scores as a Predictor of Credit Risk
FICO scores are used by almost all of the lending industry as a measure of risk for individual borrowers. Lenders are able to maximize profits by minimizing the risk they take on relative to the return they receive from the loan. The return they are getting from the loan is the interest rate the borrower pays. The lower the credit score (higher risk), the higher the interest rate and expected return. This represents the risk reward payoff of the loan. Figure 2, below, illustrates the relationship between FICO scores and negative loan performance. This evidence suggests that FICO scores are good indicators of average expected credit performance which means when all the loans are taken they work out to show a good prediction rating, but when just one loan is looked at, it may not be the case.
Figure 2. Source – The Fair Isaac Corporation.
As can be seen in Figure 2, negative performance in each bin, or score range, of FICO scores is much lower for real estate borrowers than for all consumers. This suggests that people may not have the best understanding of how their activities affect their credit. The reason some borrowers have low FICO scores in the first place may be because they do not fully understand what goes into the score. An example of this could be if a person makes mortgage payments on time but has missed student loan payments because they see them as less important. Another consideration is that when a mortgage is made by a lender, much more information is gathered about the borrower. Intuition suggests that when more information is collected, a better assessment of risk can be made. This also illustrates the point that commercial lenders must consider many other factors when making loans.
Although there are some differences in default rates between loan types, these differences tend to converge over time. Past research has shown “… that ARM default rates increase by a factor of 1.5-2.5 times, whereas fixed default rates double or triple over the extended time horizon” (Kraft 8). This research implies that the predictability of defaults comes more from FICO scores than from loan type. More recent research, however, implies the opposite.
New data from the Mortgage Bankers Association show that mortgage credit quality problems go well beyond the subprime sector. This can be seen from the fact that delinquencies on prime adjustable-rate mortgages are rising quickly – much more quickly, in fact, than those on subprime fixed-rate loans (Hatzuis 1).
Definitions of “prime” and “subprime” vary from source to source, but an acceptable cutoff is at a FICO score of 675. People with scores below 675 are classified as “subprime” borrowers and those with scores above 675 are classified as “prime” borrowers. This relationship can be seen in Figure 3 below through the increase in delinquency rates for prime and subprime, adjustable-rate and fixed-rate mortgages.
Figure 3. Source – Mortgage Bankers Association.
On average, FICO scores are a good measure of risk. This can be seen clearly in Figure 2; as FICO scores increase, negative loan performance decreases. However, Figure 3 suggests there are other important variables to consider when determining credit risk. In Figure 4 below, the default rate on adjustable-rate mortgage loans for the FICO bucket 800-850 is higher than that of the 700-719 bucket. Therefore, it is dangerous for lenders to rely too heavily on FICO scores as the sole measure of credit risk.
Figure 4. Default Rates
Source – The Fair Isaac Corporation
Issues Relating to Privacy and Technology
As the processes for developing credit risk indicators become more efficient, the speed with which lenders are able to make decisions increases dramatically. The increased use of FICO and other credit scores has helped the entire lending industry. “Especially for consumers with relatively good credit, approvals for loans can be given in a fraction of the time previously required, without any manual review of the information” (“Credit Score Accuracy” 4). It is unlikely that the recent boom in loan originations, particularly in the mortgage sector, would have been possible without the use of FICO scores (“Credit Score Accuracy”). Aside from the added speed credit scores can bring to lending, there are also the specialized products and services now available. Borrowers can shop around for the best prices on loans and lenders can create exotic loan choices to best cater to the needs of both parties. Overall, lending has become much more efficient since the advent of credit scoring, but the costs of this added benefit could be considered high.
Scores That Imitate the FICO Score
A VantageScore is a credit rating score distributed by the 3 credit reporting agencies: Equifax, Experian, and TransUnion. This score uses the same credit information but is calculated differently than a FICO score. Although this score basically measures the same types of risks, it is not directly comparable. A VantageScore is measured on a scale of 501 to 990 which is much higher then the FICO range of 300 to 850. One of the major problems with using this score is that not many lenders use it when evaluating the risk of a borrower. Problems tend to arise when someone gets what they believe to be a relatively high VantageScore, 678 for example, and thinks they will get a prime rate on a mortgage. In all actuality, the same person would likely have a much lower FICO score, possibly around 578, which would drop them into the category of a subprime borrower. “It's as if she turned on the radio, heard it was 32 degrees outside, put on a coat and stepped out, only to find that the temperature was 32 degrees Celsius -- about 90 degrees Fahrenheit” (Lewis).
The development of VantageScore was put in place so the credit bureaus would not have to share profits from releasing this score with Fair Isaac. The credit rating bureaus claim this imitation score is easier to understand and will help people be educated about their credit. "A lot of people are finding themselves in this situation -- they think they have this score and they find out the true FICO score is different," says Michael Moskowitz, President of Equity Now, a mortgage lender in New York City. Consumers get a score that is totally useless to them, and in many cases they are not even notified that this is not the score most creditors will use. While consumers are starting to become aware of the use of the VantageScore, the institutions that utilize credit scores are not planning on using it. "With billions and billions of pieces of data, it's becoming more and more refined with each transaction” (Lewis). The amount of data that are available using FICO scores makes the possibility of changing to the VantageScore almost impossible. Even though the credit bureaus are aware this scoring method is not going to be adopted by the mortgage industry, they are trying to market it to other industries. The marketing and use of this new credit score likely creates even more confusion for borrowers regarding their credit rating.
Personal Information and Privacy
Many find the idea of tracking credit histories somewhat intrusive. When surveyed, 18% called the information collected for the calculation of FICO scores an “unnecessary invasion of privacy.” 32% claimed that this collection is “necessary for efficient lending” and 48% said that it is “necessary for efficient lending but still a slight invasion of privacy.” While some may view the calculation of FICO scores as invasive, if they choose to use credit, they have no choice as to the method of tracking their credit habits.
Privacy issues go beyond just the collection of credit information for credit score calculations. In many cases, when borrowers apply for credit, the lender will request several pieces of information beyond their FICO score to accurately assess the risk. For example, a borrower applying for a mortgage would likely be asked for employment status and history, previous home ownership, assets, income, bank statements, debts, monthly obligations, etc. This information is important to lenders when making their decisions, but at the same time, borrowers may not want every company they apply for a mortgage through to have this much information about them. Of those surveyed, 87% said that they would divulge this personal information to a lender in order to receive a lower rate. While there are no laws requiring this information be given, it may be necessary if people wish to save money on loan payments.
Implications For Individual Borrowers, Primary Lenders, and the Financial Markets
Individuals should be aware of their score and how their actions affect it. A borrower in the lowest basket of FICO scores will likely pay the highest rate for a loan, if they are able to get a loan at all. There are times when it is impossible for an individual with a very low FICO score to get a loan because of laws that restrict the interest rate a lender may charge. If this is not high enough to compensate for the chance of default, the lender will refuse the credit application.
Lenders must also be aware of the mechanics of FICO scores. There are two ways a FICO score can misrepresent the underlying risk of the borrower. The first is by giving too low of a score to a particular person. Because the FICO score is understated, the lender will charge a higher interest rate to compensate for the higher perceived risk. This hurts borrowers because they will have to pay a premium for the loan. The alternative is when a FICO score is higher than it should be. This is beneficial to the borrower, but can have a negative effect on the institution that is lending the money. This higher FICO score would portray the borrower as a less risky loan candidate, charging them an interest rate that is too low for the level of actual risk involved.
Since the evidence for the statistical validity of FICO scores indicates that they are a good indication of average credit risk, institutional investors do not need to be as concerned with a specific individual’s FICO score accuracy. The securities that are backed by individual loans are grouped together in large pools. When the loans are aggregated in such a way, it creates a situation where overstating and understating inefficiencies in FICO scoring cancel out, leaving a good average indicator of the credit risk. However, institutional investors do need to be mindful of other factors that can lead to delinquency such as the type of loan, refinancing rates and other market factors.
Future Research of FICO Scores
There is still more research to be done in this area. The findings of this chapter are not conclusive as there is no way to definitely determine the accuracy of FICO scores. The survey was from a sample comprised mainly of college students. Future research should include a much wider range of ages to better estimate the entire population. Also, correlations of information such as age and gender would be useful to combine with FICO scores. As credit markets are always changing, so should the research. Issues relevant today may be less important tomorrow and the scrutinizing of processes in place should be ongoing.
Conclusions for FICO Scores as a Measure of Risk
Our research has shown that while the FICO score is a good measure of risk for the population as a whole, it does not represent risk very well on an individual level. Some people are riskier than their scores would indicate, while others are less risky than their scores show. Therefore, when considered in aggregate, FICO scores are a good indicator of credit risk. Our findings also show it is almost impossible to manipulate a FICO score due to the protection of proprietary models. Even if this manipulation were possible, an efficient market would likely pick up on this and everyone would be doing it. This would in turn push scores higher and adjustments would have to be made in the model to calculate the score based on the new set of parameters. Statistical evidence and data collected through surveys suggest that many new borrowers do not have a very strong understanding of how their actions affect their credit. If the general population were more educated on credit rating systems, the lending industry would possibly become more efficient. However, this efficiency could come at the expense of personal privacy. This returns the essential theme relating to privacy and technology, the tradeoff between privacy and convenience, cost verses benefit.
Credit Score Accuracy and Implications for Consumers. Consumer Federation of America, 2002.
Fair Isaac Corporation, Consumer Federation of America. Your Credit Scores, 2005.
Hatzuis, Jan. Mortgage Credit Quality Problems Go Well Beyond Subprime. US Daily Financial Market Comment, 2007.
Kraft, Dennis F., and Jim Anderson. Structured Products Research: Consumer ABS. Wachovia Securities, 2004: 1-15.
Lewis, Holden. “Conflicting credit scores cause confusion." 07 DEC 2006.
Bankrate.com. 21 Mar 2007 http://www.bankrate.com.
myFICO. Understanding your FICO Score, 2005.
Schrock, Jason, and Ron Kirk. Credit and Insurance Scores. Colorado Legislative Council Branch 02-09, 2002.
Back to Top