Consumer Credit - G.19

Performance Evaluation of the Consumer Credit (G.19) Statistical Release
December 2011

I. Introduction

The Federal Reserve System's G.19 Statistical Release, "Consumer Credit," is published monthly, generally on the fifth business day of the month. The release provides national aggregate amounts (by type and holder) of outstanding credit held by individuals; it also contains average terms of credit. These data are compiled by the Federal Reserve System in order to monitor household debt developments as part of the monetary policymaking process.

The top panel on the first page of the G.19 release presents growth (in simple annual percentage rate terms) and outstanding levels of seasonally adjusted consumer credit (in billions of dollars) for the most recent three months, as well as for the past five quarters and the past five years. Data for the most recent month are marked on the release as preliminary and are subject to revision. The bottom panel on the first page contains data on terms of credit: interest rates on selected types of loans at commercial banks on a quarterly basis, as well as interest rates, loan maturities, loan-to-value ratios, and average amounts for new-car loans at auto finance companies on a monthly basis¹. All of these concepts are explained in footnotes on the second page of the release. The second page of the release also provides a not-seasonally-adjusted breakdown of consumer credit by type of credit and holder.

To construct the consumer credit statistics, the Federal Reserve maintains direct data collection programs for commercial banks and finance companies and obtains information for other creditor groups through secondary sources. The Federal Reserve sources include

1. Weekly Report of Selected Assets and Liabilities of Domestically Chartered Commercial Banks and U.S. Branches and Agencies of Foreign Banks (FR 2644; OMB No. 7100-0075)
2. Domestic Finance Company Report of Consolidated Assets and Liabilities (FR 2248; OMB No. 7100-0005)
3. Quarterly Report of Interest Rates on Selected Direct Consumer Installment Loans (FR 2835; OMB No. 7100-0085)
4. Quarterly Report of Credit Card Interest Rates (FR 2835a; OMB No. 7100-0085)
5. Automobile Finance Terms (FR 2005)

Secondary sources include the Federal Financial Institutions Examination Council (FFIEC), the Office of the Comptroller of the Currency (OCC), the Credit Union National Association (CUNA), the U.S. Department of Education, the Census Bureau, and private-sector products that provide information about revolving credit at retailers and consumer credit held in securitized pools of assets.

The remainder of this document includes a more detailed discussion of the history of consumer credit, a description of the various data sources, and other information called for by Statistical Policy Directive No. 3, section 8, Performance Evaluation.

II. Background

The Federal Reserve System has been the primary producer of aggregate statistics on consumer credit since 1942. This role stems from the Federal Reserve's responsibility for monetary policy and its particular responsibilities and data needs in the consumer credit area, originally under Executive Order 8843 of August 9, 1941 and Board Regulation W - Consumer Credit, and later under various other acts and regulations. Collection of information on terms of credit began in 1972 at the request of the President's Committee on Interest and Dividends, which was established as part of phase I of President Nixon's wage-price controls program.

Since the Federal Reserve began compiling and publishing data on consumer credit, both the format of the G.19 release and the underlying data have undergone many changes. A major revision to the consumer credit statistics occurred in early 1983 as a result of an exhaustive review by the Federal Reserve of its entire consumer credit data collection and compilation program; changes were made to the program to focus on items of highest priority to the Federal Reserve's analysis, to eliminate materials of dubious quality, and to reduce reporting burdens². In 1989, estimates of consumer loans held in pools of securitized assets began to be included in the estimates of total consumer credit outstanding; this change resulted in a break in the series.

In 1999, the automobile and "Other" outstanding credit series in the G.19 release were combined into a single category called "Nonrevolving" credit. This change was implemented because the estimated split between these categories lacked a suitable benchmark. The report was also reformatted to provide additional historical data.

In 2003, student loans extended by the federal government and by SLM Corporation (the parent company of Sallie Mae) were added to the estimates of total consumer credit outstanding. The G.19 release was revised to include a new line item, "Federal government and Sallie Mae," which showed data for direct federal government student loans and student loans on Sallie Mae's balance sheet. Student loans held by the non-GSE portion of Sallie Mae's parent company, SLM Corporation, were added to the finance company sector, and Sallie Mae's securitized loans were added to pools of securitized assets. After the completion of Sallie Mae's privatization in the fourth quarter of 2004, data for SLM Corporation was entirely accounted for in the finance company sector and the line item was renamed "Federal government" in December 2009.

III. Sources and Methodology

A. Credit Stocks

1. Commercial Banks

Monthly estimates of consumer credit held by commercial banks are based on data collected by the Federal Reserve in the Weekly Report of Selected Assets and Liabilities of Domestically Chartered Commercial Banks and U.S. Branches and Agencies of Foreign Banks (FR 2644; OMB No. 7100-0075)³. The report collects, among other data, information on outstanding balances of consumer revolving and nonrevolving credit as of the close of business each Wednesday. The report also collects outstanding balances of consumer loans that have been pooled and used as collateral for asset-backed securities (ABS); these loans are no longer on the balance sheets of the originating commercial banks, but they remain consumer debt.

The reporting panel for the FR 2644 consists of an authorized stratified random sample of about 875 domestically chartered commercial banks and U.S. branches and agencies of foreign banks; participation is voluntary. The Federal Reserve Banks collect FR 2644 data from the individual respondents in their respective Districts. The Reserve Banks edit the data and then transmit the information to the Federal Reserve Board for central processing. The Board staff review the sample data and make adjustments for mergers and acquisitions, as well as for the entry and exit of respondents when constructing national estimates. These data are then reported in the Federal Reserve System's weekly H.8 Statistical Release, Assets and Liabilities of Commercial Banks in the United States.

To construct the G.19 commercial bank consumer credit estimates, month-end estimates are created from the H.8 data using weighted averages of the data from the two Wednesdays before and the two Wednesdays after each month's end. The month-to-month growth rates implied by these estimates are then used to extrapolate the previous estimates for the commercial bank universe forward. On a quarterly basis, the G.19 commercial bank consumer credit estimates are benchmarked to the quarterly Consolidated Reports of Condition and Income (Call Report) (FFIEC 031-034; OMB No. 7100-0036)⁴. The Call Report is filed by the universe of insured commercial banks.

2. Finance Companies

Estimates of consumer credit held by finance companies are based on the Federal Reserve's monthly Domestic Finance Company Report of Consolidated Assets and Liabilities (FR 2248; OMB No. 7100-0005).

The FR 2248 collects selected balance sheet items as of the close of business on the last business day of the survey month. It also collects supplemental information such as outstanding balances of pools of securitized assets that are backed by loans and leases. These loans and leases are no longer on the balance sheet of the finance companies. Certain core items are collected every month, while other items needed by the Federal Reserve on a less frequent basis are collected only for the final month of each quarter.

The particular FR 2248 items used for the G.19 release are on-balance-sheet asset items 3B(1), 3B(2), and 3B(3) (consumer motor vehicle loans, revolving consumer credit, and other consumer loans, respectively) and their off-balance-sheet counterparts, items S.3B(1), S.3B(2)2B, and S.3B(3)⁵. All of these items are collected monthly⁶.

The current reporting panel for the FR 2248 consists of a random, stratified sample of about 38 companies. Under long-standing practice, certain large companies submit their FR 2248 reports directly to the Board. Reports for other companies are collected by the Federal Reserve Banks in their respective Districts; the Reserve Banks then edit these data and transmit them to the Federal Reserve Board for central processing. At the Board, the data are reviewed and then expanded to the national finance company totals by assuming that the universe grows at the rate of the sample (with adjustments for firms entering and exiting the sector).

Every five years, the Federal Reserve conducts a larger and more detailed survey of finance companies, the Quinquennial Finance Company Questionnaire and Survey (FR 3303s; OMB No. 7100-0246). The results are used to benchmark the monthly finance company estimates constructed from the FR 2248 and to update the FR 2248 panel. The most recent benchmark survey was conducted as of December 31, 2010. Using survey data from the 38 current regular monthly reporters and a stratified random sample of additional companies, a population of finance companies is estimated. Estimates of assets and liabilities for the population are then constructed.

3. Credit Unions

Outstanding consumer installment credit held by credit unions is provided by the CUNA; these data include loans held by both federally insured and other credit unions. CUNA benchmarks its data to the National Credit Union Administration's Statement of Financial Condition (NCUA 5300; OMB No. 3133-0004).

4. Student Loans Associated with the Federal Government

The amount of outstanding student loans guaranteed or funded directly by the federal government is estimated from data provided by the Department of Education and the Department of the Treasury. The estimates are partly based on monthly data on federal direct student loans from the Department of Education, which were formerly from the Treasury Department's Monthly Treasury Statement. The estimates are also based on quarterly data on federally guaranteed student loans from the Treasury Department; prior to the receipt of these data, the monthly amounts are derived judgmentally by the Federal Reserve staff. Estimates through September 2004 (when Sallie Mae was dissolved) also include loans on Sallie Mae's balance sheet.

5. Savings Institutions

Estimates of outstanding consumer installment credit at savings institutions are based on the Office of the Comptroller of the Currency's quarterly Thrift Financial Report (OTS 1313; OMB No. 1550-0023)⁷. Prior to receiving a given quarter's data, the monthly data for that quarter are estimated judgmentally by the Federal Reserve staff. Upon receipt of the data, monthly data are revised so that the quarter-end totals match those from the report.

6. Nonfinancial Businesses

Outstanding consumer nonrevolving debt held by nonfinancial businesses (retailers) is estimated using monthly retail sales data from the Census Bureau. The Federal Reserve staff use information about the historical relationship between retail sales and nonfinancial business installment credit to create the estimates. The estimates are adjusted for changes over time in the credit programs offered by nonfinancial businesses (for example, many retailers' programs have been taken over by commercial banks and the associated receivables have been shifted to that sector's estimates).

Monthly estimates for outstanding revolving consumer debt held by nonfinancial businesses are based primarily on the typical seasonal patterns of revolving credit. The monthly estimates are benchmarked once a year to data from a private-sector publication on the outstanding revolving credit of major retailers.

7. Asset-Backed Securities

The G.19 release includes outstanding consumer credit that is being held in pools of securitized consumer loans (also known as ABS). The estimates are based, in part, on some of the sources discussed previously: The FR 2644 report captures loans securitized by commercial banks, and the FR 2248 report captures those securitized by finance companies. For loans securitized by finance companies not included in the FR 2248 and for the revolving portion of loans securitized by savings institutions, the Federal Reserve uses security-level consumer ABS data from a private vendor. For other loan securitizations, the Federal Reserve staff base their estimates on industry information about the initial balance of pools and assumptions about how quickly the loans are paid down.

B. Terms of Credit

1. Commercial Banks

The Federal Reserve collects quarterly data on interest rates for 48-month and 60-month new-car loans and 24-month personal loans through the Quarterly Report of Interest Rates on Selected Direct Consumer Installment Loans (FR 2835; OMB No. 7100-0085). The data represent the "most common rates" (based on dollar-volume terms) for loans made as of the Saturday ending the first full week in February, May, August, and November⁸. For variable-rate loans, the banks are asked to consider the initial interest rate on the loans.

In recent quarters, about 125 commercial banks have filed FR 2835 reports. Banks report to the Federal Reserve Banks in their respective Districts. The Reserve Banks edit the information and then transmit it to the Board for central processing. The Board staff review the data and calculate averages that are published as part of the G.19 release⁹. These data are neither benchmarked nor adjusted for seasonal variation.

In a separate report, the Quarterly Report of Credit Card Interest Rates (FR 2835a; OMB No. 7100-0085), the Federal Reserve collects quarterly data from commercial banks on the average interest rate across all of their credit card accounts, finance charges on credit card accounts, total number of credit card accounts, total average daily balances, and total average daily balances that incurred a finance charge¹⁰. These data are collected for the billing cycles in February, May, August, and November.

The current reporting panel for the FR 2835a consists of 41 commercial banks. Respondents send their reports to the Federal Reserve Banks in their respective Districts. The Reserve Banks edit the information and then transmit it to the Board for central processing. The Board staff review the data and calculate two interest rate measures for the G.19 release. The measure labeled "All accounts" is a weighted average of the interest rate for all accounts, where each bank's weight is based on its number of accounts. The measure labeled "Accounts assessed interest" is calculated by dividing total finance charges by total average daily balances that incurred finance charges. This latter measure is designed to correspond with the average interest for cardholders who actually paid interest. These data are neither benchmarked nor adjusted for seasonal variation.

2. Finance Companies

The estimates of terms of automobile credit at finance companies are based on the FR 2005 report, Automobile Finance Terms, which collects monthly data on vehicle loan interest rates and nonrate terms from finance companies associated with automobile manufacturers. Respondents send their reports directly to the Federal Reserve Board; staff members then review the data and calculate volume-weighted averages for publication in the G.19 release.

IV. The G.19 Release

A. Processing and Publication Schedule

The Federal Reserve aims to make the G.19 data available to the public in the afternoon of the fifth business day of the second month following the reference month¹¹. This publication lag exceeds the goal stated in Statistical Policy Directive No. 3 (release within 22 working days of the end of the reference period), but it is the shortest feasible lag given the time required to collect, process, and review preliminary information from the various sources and compute the estimates. Prior to November 1986, the G.19 release had been scheduled for "about the middle" of the second month following the reference month. Streamlining of the Federal Reserve's production processes allowed the schedule to be accelerated to the current timing in November 1986. Since November 1986, the targeted release date has been met on all occasions.

Release of the G.19 data is handled by the Federal Reserve Board's Public Affairs Office. The release is available on the Federal Reserve Board's public website¹².

B. Avoidance of Premature Disclosure

All data in the G.19 release are provided to the public in aggregate form only. With the exception of commercial bank interest rate data discussed previously, individual respondent data are not made available.

Prior to publication each month, access to the estimates is limited to Board staff members who are responsible for preparing the estimates and the G.19 release. In addition, some of the information may be made available to members of the Board and the Federal Open Market Committee as background material for making monetary policy decisions (and to the staff who prepare such materials).

There is no record of a premature disclosure during the evaluation period.

C. Accuracy and Reliability

Monthly growth rates for the G.19 consumer credit series can be quite volatile. For example, the standard deviation of the annualized percent change in seasonally adjusted monthly consumer credit over the past three years is 3.7 percentage points for total consumer credit, 5.0 percentage points for revolving credit, and 3.9 percentage points for nonrevolving credit. It is difficult to assess how much of this volatility arises because of true variation in the amount households have borrowed and how much stems from error in measurement. However, the latter factor seems likely to have been at least somewhat important given the degree to which source data are incomplete at high frequencies.

The G.19 data are revised on a regular basis. Revisions to the commercial bank estimates most commonly result from respondents to the FR 2644 report reclassifying loans and making other corrections to their data and from the quarterly benchmarking to the commercial bank Call Report. Revisions to the finance company estimates generally stem from respondents to the FR 2248 report correcting their data or providing their data late, and from the quinquennial benchmarking to the FR 3033 survey. Revisions to the credit union estimates occur because CUNA benchmarks its data to information from the NCUA. Revisions to the estimates for savings institutions, nonfinancial businesses, and student loans occur because full monthly source data are not available and the staff make judgmental estimates of the likely trajectory prior to the release of the data (quarterly for savings institutions and some components of student loans and annually for the revolving component for nonfinancial businesses). Revisions to securitized consumer credit most frequently occur because the components based on commercial bank and finance company survey data are revised to reflect new information from respondents. In addition, the re-estimation of seasonal factors once a year affects the seasonally adjusted totals for overall consumer credit, revolving credit, and nonrevolving credit shown in the release.

Revisions may also occur because the Federal Reserve staff occasionally alter the methods used to estimate outstanding consumer credit. These changes may be spurred by the availability of new data sources, the discontinuation of prior data sources, or changes in loan markets that render the existing methods inferior or incomplete.

In practice, revisions to monthly growth rates can be sizable. Over the past three years, for example, the standard deviation of the difference between the initially released estimate of the annualized percent change in monthly seasonally adjusted consumer credit and the current estimate has been about 2.2 percentage points for total consumer credit, 3.2 percentage points for revolving credit, and 2.8 percentage points for nonrevolving credit. These figures are fairly large compared with the average annualized growth rates of negative 1.6 percent, negative 6.7 percent, and 1.2 percent for the three categories. In general, much of the revision for a given month's estimates occurs in the first few months after their initial release; the standard deviations of the difference between the initially released estimates of monthly growth and the estimates following the benchmarks to commercial bank and savings institution Call Reports a few months later are 2.1 percentage points, 2.4 percentage points, and 2.6 percentage points for total, revolving, and nonrevolving credit, respectively.

D. Dissemination of Revised and Historical Data

The release tables contain markings that indicate whether the data for a given time period have been revised. In addition, footnotes are added to explain some larger revisions and, for particularly important revisions, supplementary background material has been posted on the Board's public website¹³. The Federal Reserve's experience has been that the conceptual explanations provided in the footnotes and supplementary material are sufficient to inform the public. Generally speaking, many calls result from users either not reading the material or from users getting information from a secondary source such as a newspaper or newsletter. On rare occasions, outside researchers call to clarify a minor technical point with Federal Reserve staff. The Federal Reserve is sensitive to the issue of clarity and, should the number of calls concerning a specific point become large, the point would be addressed through additional footnotes or material on the website; as yet, such augmentation has not been needed.

The Federal Reserve Board makes current estimates of historical data on consumer credit available on the Board's public website.