The Quality of Credit Report Data

The lead article in the February issue of the Federal Reserve Bulletin is a masterful overview of consumer data and credit reporting. It was written by three economists from the Board's Division of Research and Statistics (Robert B. Avery, Paul S. Calem, and Glenn B. Canner) and one at the University of Southern California (Raphael W. Bostic). Since their article covered 27 pages, this review can only skim the findings as a means of providing some insight into the scope and quality of the study. While you cannot purchase the Bulletin at your local newsstand, you can access it at: www.federalreserve.gov/pubs/bulletin/2003/03bulletin.

This study is the first step in a major research effort of the staff of the Federal Reserve Board "to obtain more detailed and timely information on the debt status, loan payment behavior and overall credit quality of U.S. consumers." To gather the relevant data, the Federal Reserve Board obtained the cooperation of one of the three large credit reporting companies to supply from their credit records a nationally representative sample of individuals. The records for each person in the sample were voluminous and included current and past debts, payment history, public records relevant to each consumer's credit standing (e.g. bankruptcy filings and collection agency accounts) and credit inquiries.

Credit reports play a key role in facilitating the granting of consumer credit. The authors note that, "Industry sources report that each of the three credit reporting companies issue approximately two million consumer credit reports each day." In addition, each agency receives more than two billion items of information during the course of each month and typically enters this information in their credit records within seven days. The authors provide detailed information on the characteristics of the accounts in their sample. Having examined the data provided in credit reports, they conclude: "Overall, research and creditor experience has consistently indicated that credit reporting company information, despite any limitations that it may have, generally provides an effective measure of the relative credit risk posed by prospective borrowers."

The data base

The national credit reporting company provided the researchers with about 248,000 account records containing 2.58 million credit accounts. The authors estimate that the sample is representative of the approximately 1.43 billion accounts contained in the bureau's full national database, totaling $6.7 trillion as of mid-1999. All personally identifiable information was removed prior to the author's receipt of the data. The account records provided a massive amount of data: over 350 variables describe account dates (date opened or closed), account balances, payment performance, account description, loan purpose or type, lender subscriber number.

While closed-end credit dominated the dollar amount that consumers owed, open-end credit dominated the number of accounts held by consumers, accounting for 64.3 percent of all accounts. Mortgages made up two-thirds of the dollar amount owed on closed-end credit, but only 5 percent of the number of accounts. All revolving credit accounts had an average balance of $2,015, but the average amount outstanding ranged from $378 at retailers to $9,736 at banking institutions.

It is interesting to note that 22.9 percent of all revolving credit accounts were open, but with no balance. Not unexpectedly, about 30 percent of retail revolving credit accounts had no balance at the time of the survey. The data show that in their own self-interest, creditors are careful and consumers, responsible. The authors point out that most delinquencies are temporary. Consumers are delinquent for a month or so, but then move back to a current status. For example, only 0.6 per percent of revolving credit accounts were 90 to 149 days past due at the time of the survey. Just 2.2 percent of nonrevolving accounts were similarly past due.

There is much more in this comprehensive article. The authors note some problems with the quality of certain data items: "a close examination of credit reporting company data reveals that the information is not complete, may contain duplications, and at times contains ambiguities about the credit histories of at least some consumers." They list four problem areas: 1) credit limits are sometimes not reported, 2) the current status of accounts that show positive balances but are not currently reported is ambiguous, 3) some creditors fail to report nonderogatory accounts or minor delinquencies, and 4) the reporting of data on collection agency and public record accounts is possibly inconsistent and inquiry data is incomplete.

Nevertheless, we quote their last paragraph in which the authors summarized their basic conclusions:

"First, although some problems in the credit reporting data that are likely to affect the credit evaluation of individuals have been identified, it is very difficult to determine the extent to which credit availability would change if the problems were addressed. It is likely that the data issues will materially affect the availability and price of credit only for those individuals of marginal creditworthiness. Second, the costs of correcting these identified data problems have not been evaluated. Some of the problems may be very difficult and expensive to overcome, and in some cases the costs may exceed the benefits. Finally, this analysis rests on the experience of only one of the three national credit-reporting companies and uses data that are now somewhat dated. Many changes are taking place in the credit reporting industry, and they may mitigate some or all of the highlighted limitations.