Consumer Credit Data

The consumer credit statistics presented here were calculated by the Federal Reserve Bank of Minneapolis’s Community Development Department using data from the Federal Reserve Bank of New York/Equifax Consumer Credit Panel data set (consumer credit data set). This is an anonymous, nationally representative sample consisting of 5 percent of U.S. residents with a Social Security Number and a credit report.1 The data set is an unbalanced panel, meaning that certain randomly selected consumers enter the data set when their credit report is initially created, and others disappear after they die or because their credit bureau file no longer contains sufficient information.2 The data are reported quarterly and generally reflect conditions as of the last month of each quarter.

In total, the consumer credit data set includes roughly 12 million consumers. Roughly 350,000 of them are located in the Ninth Federal Reserve District, which is composed of Montana, North and South Dakota, Minnesota, northwestern Wisconsin, and the Upper Peninsula of Michigan. Both nationwide and Ninth District statistics are based on the entirety of the 5 percent sample.

Methodology for Developing Estimates for Neighborhood Income Categories

No information about the income of individuals is included in the consumer credit data set. However, the data set does include the census tract, county, and state where the consumer lives. In this analysis, a consumer is assigned to a neighborhood income category based on the median family income (MFI) of the census tract where he or she lives relative to the median family income of either the metropolitan area (for metropolitan area residents) or the nonmetropolitan portion of the state (for nonmetropolitan residents).3 The neighborhood income categories used in this analysis are low income (<50 percent of MFI); moderate income (50-79 percent of MFI); middle income (80-119 percent of MFI); and upper income (≥120 percent of MFI). The analysis presents statistics based on the credit characteristics of consumers in each of these neighborhood income categories nationally and within the Ninth Federal Reserve District.

Our methodology for classifying census tracts into income categories is based on, but not identical to, methods used by the Federal Financial Institutions Examination Council (FFIEC) in assigning census tracts to income categories for Home Mortgage Disclosure Act and Community Reinvestment Act data-reporting purposes. Rather than using more current income data from the American Community Survey (ACS) as the FFIEC does, we use the ACS data set from 2005-2009. Although these income data are older, the 2005-2009 file uses census tract geographic definitions from Census 2000, which is consistent with the information reported in the consumer credit data set. The newer ACS data used by the FFIEC are based on census tract definitions from the 2010 Census, which are not entirely compatible with those used in the consumer credit data set.

In this analysis, we exclude consumers with a nonresidential address (e.g., a PO Box) from our income category estimates but include them in all total estimates. We treat consumers with incomplete or inaccurate geographic data similarly, as well as those who live in a census tract for which a median family income estimate is not available. Consumers who have exited the sample after one year or less are excluded from the analysis entirely, as are those reported to be deceased.4

Interpreting Estimates

This section describes the consumer credit statistics calculated by Community Development Department staff. If any questions remain, please feel free to contact the department directly at

Median Credit Score
The credit risk score provided by Equifax (the Equifax Risk Score) ranges from 280 to 850, with a lower score indicating a higher level of risk. However, Equifax does not calculate a risk score for consumers whose credit histories are too limited (“thin”) to provide sufficient information for gauging credit risk. These thin-file consumers are excluded from the calculation of the median credit score, which is the midpoint of existing scores. That is, half of the consumers with scores have a score above the median and half have a score below the median.

Percent of Consumers with Poor/Fair Credit
This is the number of consumers with a score below 660, as a percent of all consumers in the sample who have a credit score. The threshold of 660 was chosen because this score is commonly used to distinguish prime borrowers from subprime borrowers.5 A consumer with a score below 660 likely encounters challenges in accessing credit from mainstream financial institutions.

Percent of Consumers in Foreclosure
This is the number of consumers who were in some stage of a mortgage foreclosure process during the quarter, as a percentage of all consumers with mortgage debt (where mortgage debt includes home installment loans but not home equity lines of credit). Being in a mortgage foreclosure process includes having a foreclosure process begin during the quarter, having a previously existing foreclosure process end during the quarter, and having a previously existing foreclosure process continue through the quarter.

Percent of Consumers with a Seriously Delinquent Tradeline6
This is the number of consumers with a seriously delinquent tradeline, as a percent of all consumers with a tradeline, where seriously delinquent includes loans that are 90+ days past due, in collections, classified as severely derogatory, or in bankruptcy. Accounts that are severely derogatory are more than 120 days past due and may be charged off by the lender as bad debt. Collection accounts reported by collection agencies or debt buyers are not included in this estimate. Consumers with no debt or an active account in good standing are considered to be current and are thus included in the denominator of this calculation.

The delinquency rate estimates presented in this analysis differ from those typically provided in other sources in at least two ways. First, these rates reflect the percent of individuals with a delinquent account. This methodology can produce very different trends than alternative specifications that calculate delinquency as the percent of outstanding debt (or loans) past due. Second, the delinquency rates presented here include individuals with accounts that lenders have likely closed and charged off because they no longer expect repayment. Many sources of delinquency rates on consumer loans will report a charged-off account in the period in which it occurs (a month, quarter, or year) but not thereafter.7 In the consumer credit data used here, seriously derogatory accounts (including ones charged off) can be reported for up to seven years. To the extent that lenders continue to report these accounts, they will be reflected in these charts.

It is worth noting that the percent of consumers with a delinquent bank card tradeline increased sharply in the last quarter of 2010 and the first quarter of 2011. At the same time, the percent of consumers with a bank card tradeline—delinquent or not—also increased, but to a lesser degree. A significant portion of the bank card tradelines that first entered the data set during this period were already reported as 120 days or more past due or in collections. This is unusual and warrants further investigation.

Percent of Consumers with Debt
For total debt, this estimate reflects the percent of consumers in the data set with at least one debt greater than $0 (excluding accounts in bankruptcy but including debt from closed accounts that the lender has charged off but continues to report to the credit bureau). The percentages by type of debt are the percent of consumers in the dataset who have at least one debt of that type with a current balance greater than $0.

Median Debt
For consumers with debt, median total debt is the midpoint of consumers’ debt totals, excluding accounts in bankruptcy but including debt from closed accounts that the lender has charged off but continues to report to the credit bureau. That is, half of the consumers have total debts that exceed the median and half have total debts that are less than the median. It is important to note that consumers with zero debt are not included in this estimate. Median debt by type is calculated only for consumers who have nonzero debt of that type. It again represents the midpoint, so that half of the consumers with that type of debt have balances greater than the median. When a credit file contains a joint debt, half of the amount owed is attributed to the consumer in question. That is, to mitigate potential double counting of debt amounts for joint debtors, it is assumed that the other half of the debt is owed by the consumer’s spouse or other partner.

Additional Resources Related to Consumer Credit Data

Robert B. Avery, Paul S. Calem, and Glenn B. Canner, “Credit Report Accuracy and Access to Credit,” Federal Reserve Bulletin (Summer 2004), pp. 297-322.

John M. Barron, Gregory Elliehausen, and Michael E. Staten, “Monitoring the Household Sector with Aggregate Credit Bureau Data,” Business Economics (January 2000), pp. 63-76.

Meta Brown, Andrew Haughwout, Donghoon Lee, and Wilbert van der Klaauw, Do We Know What We Owe? A Comparison of Borrower- and Lender-Reported Consumer Debt, Federal Reserve Bank of New York Staff Report 523 (2013).

Julia S. Cheney, Alternative Data and Its Use in Credit Scoring Thin- and No-File Consumers, Federal Reserve Bank of Philadelphia Payment Cards Center Discussion Paper (2008).

Julia S. Cheney, Payments, Credit, and Savings: The Experience for LMI Households, Federal Reserve Bank of Philadelphia Payment Cards Center Conference Summary (2007).

Robert M. Hunt, A Century of Consumer Credit Reporting in America, Federal Reserve Bank of Philadelphia Working Paper 05-13 (2005).

Andrew Kish, Perspectives on Recent Trends in Consumer Debt, Federal Reserve Bank of Philadelphia Discussion Paper (2006).

Donghoon Lee and Wilbert van der Klaauw, An Introduction to the FRBNY Consumer Credit Panel, Federal Reserve Bank of New York Staff Report 479 (2010).


1 The 5 percent sample is drawn using the last two digits of a consumer’s Social Security Number; consumers whose Social Security Number ends in one of five two-digit combinations are selected into the sample. However, the data set used to develop these statistics does not include the Social Security Number of the consumer or other identifying information (e.g., consumer name, lender name, address, etc.).

2 In order to be included in the data set, an individual must have at least one of the following in his or her credit bureau file at Equifax: an item of public record (e.g., a judgment) within the last seven years or bankruptcy filing within the last ten years; an open credit account that is regularly updated by the lender or servicer; or a closed account that continues to be reported, which can occur for up to seven years if the account was not in good standing. Not all adults will have a credit bureau file. According to an analysis of the Survey of Consumer Finances, slightly more than 8 percent of households do not include a member with a credit report (Brown, et al. 2011).

3 Income data are from the five-year American Community Survey (ACS) data set covering the years 2005-2009. In this analysis, a census tract is assigned to only one neighborhood income category for the duration of the study period, based on its income during the five years covered by the ACS data. For consistency, the same metropolitan area definitions are used throughout the study period. We use the definitions adopted for the release of the 2005-2009 ACS data and defined in the November 2008 OMB Bulletin No. 09-01.

4 It is impossible to know whether consumers who are first available in the five most recent quarters will remain in the sample for more than one year or whether they will exit after a year or less. We retain all consumers unless it is evident that they have exited the sample after one year or less. As a result, some recently added consumers included in this analysis will, in all likelihood, exit the sample as new quarters of data become available. In the eight quarters of data available in 2010 and 2011, anywhere from 17 to 35 percent of sample entrants subsequently exited the sample after one year or less.

5 See, for example, Office of the Comptroller of the Currency, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation, and Office of Thrift Supervision, Supervision and Regulation Letter 01-4(GEN): Guidance on Supervision of Subprime Lending, January 31, 2001. Accessed July 17, 2012.

6 A tradeline refers to an individual borrower-lender relationship, such as a single mortgage or a single student loan. Because lenders are permitted—but not required—to report seriously delinquent and subsequently closed tradelines and their charged-off balances to the credit bureau for up to seven years, such accounts can be included in the count of tradelines. It is important to note that the presence of a tradeline is not synonymous with the existence of debt, particularly for revolving accounts such as bank cards, for which it is common to have a tradeline but no outstanding debt associated with it. Mortgage estimates include both regular mortgages and home equity installment loans.

7 In financial statements, a charge-off is the removal of a loan (an asset) from the lender’s balance sheet. In credit bureau data, an account charge-off reflects negatively on a consumer’s repayment history and can be legally included in a credit report for up to seven years.

Latest Articles