To understand credit scores, it is first necessary to understand the primary sources of all credit information in the United States: Virtually all credit data are accessed through the three repositories, Trans Union, TRW (also called Experian), and Equifax. These companies compile information on your (and practically everybody else's) credit history which is provided to them by the creditors themselves. By far most of the data are transmitted by computer "tapes", which are then examined by these credit repositories for accuracy, etc. and, theoretically, anyway, "corrected". 

The primary areas covered in the files are:  Name of creditor, date account opened, date information reported by creditor, original amount or credit limit, current balance, nature of the account - e.g. joint or individual, installment, 30 day, or revolving,  number and dates of any late payments, and any past due amounts. The account numbers are supposedly included, but, far more often than not, they are incomplete, i.e. usually the report only reflects part of the account number. In theory, the credit file contains information which can accurately and completely summarize a person's credit profile - to whom that person has owed how much for how long and whether the creditor got paid as agreed (To digress briefly, this phrase, "paid as agreed" is the single most important concept to grasp in understanding credit; it is, for example, why paying early will never offset paying late - because paying early one month and late the next was not as agreed).

Credit scores are numbers, theoretically between 400 and 850, which are computer-generated from the data contained in a person's credit file; the higher the number, the better the score. What makes a higher score better is that it indicates a lower degree of what is known as default risk. Put another way, the lower the number, the greater the chance that the person will not pay his/her bills on time, if at all.

All three of the credit repositories have created computer programs to calculate a score. The number  which prevails - the one which the mortgage lender considers the score -  when a credit report is generated from TRW, Trans Union, and Equifax, is the score in the middle, known as the median. For example, if the numbers were 680, 701, and 710, then 701 is noted in the file as the borrower's credit score. If there are two or more borrowers, then the lesser (or lowest with 3+ applicants) is used. So, if John Doe's score is 690, and Jane Doe's is 701, then 690 prevails.

Now, more precisely what is poor, fair, good, and very good? First, it should be stated that the theoretical limits mentioned above are just that, theoretical. I am not sure any real person exists with credit so good as to have 900, or so poor as to have earned a 400. In our experience, about 1% have a score of 800 or more, and almost none under 500. For this company, roughly 1/3 are less than 660, 1/3 between 660 and 720, and 1/3 at 720 or more. In general, this is how the scores are rated by FNMA:

• 720+           Low risk                     Note: Other institutions/lenders have different

• 660-719      Moderately low  risk             demarcations; see 3rd paragraph below.

• 620-659      Moderately high risk

• 619 or less   High risk

The limits are not quite as abrupt as it might at first appear, meaning as a score approaches the next level, for example 705, the risk is regarded as less than, say, 660. Having said that, there are times, such as mortgage insurance computations, and program eligibility where the threshold is just as precise (and hence abrupt) as a grade in school (a 84.4 is a C, an 85 a B). For example, someone with a credit score of 759 might pay $73/mo for private mortgage insurance, and another with a 760 would pay $66.  Similarly, but with greater consequences, one individual who is otherwise equal (income, debts, etc) to another might be approved with a 620 and the other with 619 rejected. Most of the time the consequences of credit scores have to do with the amount one can borrow, and the interest rate one pays. 

On the issue of the effect on interest rate, the disparity can be significant, particularly if the borrower is forced to seek out what is commonly called "sub-prime" (also designated " non-conforming, high risk, graded paper, and alternative" financing). In some cases, the score might result in a rate of only 1/4% higher (roughly $17/mo for $100,000), but in the sub-prime category the rate can be 2% to 5% - or more - higher (For the record, Baker and Lindsey does not make sub-prime loans), especially for home equity and other types of second mortgage loans.

As testimony to Congress in 9/00 illustrates, rates can vary considerably (and abruptly) for car loans:

        720+  7.5%    680-719  9.5%   660-679  10.5%   500-659  17.5%

Note that the demarcations are different than FNMA's for mortgage loans, and the rate disparity is considerable.

Finally, it is useful to have some idea of how the scores are determined. Contrary to popular opinion, the absence of late payments does not necessarily mean the score will be high (I can remember an instance where there were absolutely no late pays, but the score was 580.); there is much more to the determination of the number than just the payment history.

There are basically five factors which comprise the score: 

• Payment History (35%)

• Amounts Owed  (30%)

• Length of Credit History (15%)

• New Credit  (10%)

• Types of Credit Used (10%)

Payment history is the easiest to follow; simply put, it means how many, if any, late payments have been made, and when (i.e. how recently late ). 

The term "amounts owed", more than anything else, has to do with the balances of revolving accounts as they compare to the credit limit (Revolving accounts are those where the balance may increase, as well as decrease, such as credit cards and lines of credit.). If a person had five cards with limits of $5,000 and balances of $500 each and were to consolidate those cards into one of $2,500 with a limit of that amount, his/her score would almost certainly be hurt (In more ways than one - often creditors do not report when paid off, so it would be possible for the credit report to show not only the new card with $2500 balance, but also the six cards just paid off.)

Length of credit history means just that: How long has one been in the system and the dates of the adverse data. More recent late payments will hurt the score more than older ones. Since the computer algorithms regard the unknown to be pretty much the same as recent, then the missing dates are treated the same as recent. So, if you have a late payment from a few years ago and the creditor has not reported the dates, it would be worth the effort to induce that creditor to report when you were late to at least have a shot at improving your score.

New credit refers to the number of newly opened accounts, how long it has been since the last account was opened, and the number of inquiries by potential creditors. Many consumers are aware now that too many checks into one's credit can reduce the score, but it is not always easy to know when your credit is being checked, or how often (For example, applying for car or homeowner's insurance will usually result in a credit inquiry. Theoretically, the repositories (TRW, Trans Union, and Equifax) lump several inquiries of a few days into one (i.e. as though only one inquiry occurred). Personally, I am not convinced....yet.

Types of credit refer to the nature of the account itself, i.e. whether it is an installment or mortgage loan, retail or credit card, or with a finance company (Too many finance companies will adversely affect the score, although apparently not a great deal). It has been my observation that installment loans are more likely to help the score, whereas the reverse is true for revolving accounts.

The most important point I think to make about  credit scores is that paying on time (or not) is only a part of the score, so one can have a perfect pay history and yet have a low score -  remember the one individual whose score was about 580,  who had not a single late payment. He did have all credit cards, each with high balances at or near the limits, and all the accounts were no more than two years old at the time of the report. Thus, when I bring up the subject of credit scores and am told "I have good credit" the person may be telling me the truth, in that he has not been late on any payments, but his/her credit score may still be low enough to place that person in a high risk category. The score is not something a consumer is likely to be able to intuit accurately, at least not yet, so the only way to find out for sure is to obtain a credit report which also reflects the credit scores (something not generally available to the consumer - contact a lender to get this information).