Board of Governors of the Federal Reserve System

June 14, 2024

A Note On Revolving Credit Estimates¹

Michael M. Chernousov, Jessica N. Flagg, Simona M. Hannon, Virginia L. Lewis, Suzanna M. Stephens, and Alice H. Volz^*

Revolving credit represents a notable share of consumer debt and is an important part of household balance sheets. At the end of 2023, revolving credit was measured at over $1.3 trillion in the Z.1 Statistical Release, "Financial Accounts of the United States," and accounted for more than 25 percent of total consumer credit.² (With the exception of the pandemic period, this share has hovered above 25 percent over the past decade.) These estimates are based on the G.19 Statistical Release, "Consumer Credit," produced by the Federal Reserve Board and are included in the set of Principal Federal Economic Indicators. That said, several other data sources can be used to construct estimates of revolving credit. In this note, we present an overview of the construction of the revolving credit estimates in the G.19 and compare these estimates to alternative measures of consumer credit.

Definition and Data Sources

Revolving credit primarily consists of credit card debt, but it also includes other revolving credit plans linked to overdraft arrangements and other revolving credit plans not accessed by credit cards. As of the end of 2023, data from the Reports of Condition and Income (Call Reports) for the commercial banking sector—the largest holder of such credit—indicate that approximately 92 percent of revolving credit balances consist of credit cards, with the remaining 8 percent attributed to revolving credit plans. Although commercial banks hold nearly 80 percent of revolving credit in the G.19 data, the entire universe of institutions holding such balances includes thrifts, credit unions, nonfinancial businesses, and finance companies.

Estimates of revolving credit in the U.S. can be obtained from five data sources. As summarized in Table 1, these sources vary in their representation of the revolving credit universe and in the frequency at which they are collected.

Table 1: Revolving Credit Data Source Overview

1. The G.19 Statistical Release, "Consumer Credit," reports outstanding credit extended to individuals for household, family, and other personal expenditures, excluding loans secured by real estate.³ The G.19 provides seasonally adjusted data for consumer credit outstanding, as well as simple annualized percent changes and not seasonally adjusted levels of total, revolving, and nonrevolving credit. Data are collected by sector—banks and thrifts grouped together under the depository institutions category, finance companies, credit unions, the federal government, non-profit and educational institutions, and nonfinancial businesses—from various sources that include weekly and monthly submissions to the Federal Reserve Board, as well as quarterly Call Report information. The G.19 covers the entire universe of consumer credit and is published at a monthly frequency.
2. The H.8 Statistical Release, "Assets and Liabilities of Commercial Banks in the United States," provides an estimated, weekly aggregate balance sheet for all commercial banks in the United States.⁴ The H.8 release is primarily based on data that are reported weekly by an authorized sample of 850 domestically chartered banks and foreign-related institutions and quarterly Call Report data. Line item 21 in Table 2, "Credit Cards and Other Revolving Plans", presents estimates of revolving credit at domestically chartered commercial banks in the U.S.⁵ Like the G.19, the H.8 shows seasonally and not seasonally adjusted levels and growth rates based on season- ally adjusted levels for revolving credit. The H.8 covers the commercial banking universe and is published weekly.
3. Estimates of revolving credit in the U.S. can also be obtained using Federal Reserve Bank of New York Consumer Credit Panel (CCP)/Equifax data. The CCP is a database on consumers' credit use and payment performance drawn from an anonymized 5 percent random sample of Equifax credit bureau records.⁶ Estimates based on the quarterly CCP version available to us are considered representative for the entire consumer credit universe in the U.S. reflected in credit bureau data.
4. Additional estimates of revolving credit can be obtained from the Survey of Consumer Finances (SCF), which is a survey produced by the Federal Reserve Board in partnership with the Department of the Treasury.⁷ The SCF is a triennial, cross-sectional survey of U.S. families. The survey includes information on families' balance sheets, pensions, income, and demographic characteristics. The SCF contains an oversample of wealthy households, ensuring coverage of the full distribution of household wealth. Estimates based on the SCF are considered representative for the universe of revolving debt held by U.S. households.⁸
5. Finally, the Capital Assessments and Stress Testing (FR Y-14M) data reflect a large share of revolving credit balances in the U.S. The data are collected by the Federal Reserve Board, while the estimates discussed are produced and publicly made available by the Federal Reserve Bank of Philadelphia.⁹ The FR Y-14M report collects detailed monthly domestic credit card data (among others). The respondent panel is comprised of U.S. bank holding companies, U.S. intermediate holding companies of foreign banking organizations, and covered savings and loan holding companies with $100 billion or more in total consolidated assets.¹⁰ As a result, the universe coverage is imperfect because the data collection does not include smaller lenders. The data have a monthly frequency.

Comparison of Estimates across Data Sources

Looking first at two related statistical releases produced by the Federal Reserve Board that publish levels and growth rates for revolving credit—G.19 "Consumer Credit" and H.8 "Assets and Liabilities of Commercial Banks in the United States"—we note that the monthly G.19 Statistical Release aims for complete coverage of the consumer credit universe in the U.S., while the weekly H.8 covers only the commercial banking sector. Since commercial banks are the major holder of the outstanding credit card debt, the principal data input for the G.19 revolving consumer credit series is from the weekly H.8 Statistical Release, line item 21, "Credit Cards and Other Revolving Plans."

Beyond differences in data coverage, there are several additional reasons why the measures of the G.19 revolving consumer credit and the H.8 credit card loans and other revolving plans series can differ in a given month. The second potential source of discrepancy arises from the frequency conversion used to calculate monthly levels of revolving consumer credit. The G.19 release uses the not seasonally adjusted, weekly H.8 data on credit card and other revolving plans and converts them into a monthly series by estimating the end-of-month levels. By contrast, the H.8 release calculates the pro rata averages of the Wednesday levels to construct the monthly level series. These computational differences can lead to considerable differences in the percent changes (that is, growth rates) between the G.19 and H.8 releases, especially in months with large changes in revolving credit card debt between the beginning and the end of the month.¹¹

Differences can also arise as a result of seasonal adjustment. Both the H.8 and G.19 releases publish seasonally adjusted level series. However, there are two key differences in the methodologies employed. The H.8 release uses a multiplicative seasonal adjustment process to create seasonally adjusted weekly level series and then uses those series to construct seasonally adjusted monthly average levels and percent changes. The G.19 release, however, uses an additive process to seasonally adjust monthly flow values and then recursively constructs the seasonally adjusted level series and percent changes using the seasonally adjusted monthly flows. Additional differences in seasonal adjustment model selection can also cause discrepancies between the two measures.

Adjustments for structural activity in the H.8 release can further contribute to differences, creating a unique cause of discrepancy between the final published measures of the percent changes. In order to calculate the percent changes, both releases calculate flows of the series. However, for the percent changes, the H.8 release uses nonpublic level series that include adjustments for nonbank structure activities and accounting rule changes (e.g., a commercial bank acquires a thrift or adopts FAS 166/167), while the G.19 release incorporates the level series publicly reported on the H.8 release.¹²

Another source of discrepancy arises because of quarterly scaling of the monthly G.19 depository institutions estimates to the Call Report. H.8 estimates cannot be used directly, as the H.8 release is restricted to domestically chartered commercial banks which excludes savings institutions (e.g., thrifts) and some other depository institutions. To capture the entire universe of depository institutions, the G.19 release uses the quarterly Call Reports, which include these other institutions, as a benchmark and scales up the monthly data obtained from the H.8.

Finally, the G.19 release also adds adjustments, such as those related to interbank activity or transfers of loans between financial sectors, which may add to the discrepancies.

Although the factors discussed above contribute to differences between the revolving credit levels and growth rates reported by each release, the growth rates are very similar. In panel (a) of Figure 1, we show the G.19 and H.8 monthly annualized growth rates over the past decade. In panel (b) of Figure 1, we compare all revolving credit data sources discussed in this note. As can be seen in the figure, the SCF provides the lowest estimates of credit card debt, while the G.19 estimates are the highest, followed by estimates constructed from credit bureau data (CCP).

Figure 1. Revolving Credit Growth Rates and Outstanding Balances

Unlike the H.8, which focuses on commercial banks, the SCF, G.19, and CCP aim to reflect the entire universe of revolving credit; however, there are differences between the data sources. First, there are differences related to the timing of the data collection. As previously noted, to estimate the month-end level, the G.19 release uses the not seasonally adjusted weekly "Credit Card and Other Revolving Plans" H.8 data and converts them into a monthly series. In addition, the G.19 revolving credit holdings by nonbank sectors are at varying frequencies: Thrifts, quarterly; credit unions and finance companies have holdings reported as of the last day of the month; whereas estimates for nonfinancial business are constructed based on an annual benchmark. The CCP, however, provides estimates of revolving credit at the end of each quarter. In addition, the CCP-derived estimates can be subject to delays in loan servicer reporting, which do not affect other data sources. Furthermore, whereas the G.19 and the CCP measure balances at a discrete point in time without reference to the payment cycle, the SCF asks about credit card balances as of the time the respondent made their last payment and thus excludes charges incurred in the interim. The measurement of revolving credit in the triennial SCF is also subject to some drawbacks, such as the lag between data collection and data release and the relatively small sample size (Henriques and Hsu, 2013).

Other discrepancies can be attributed to differences in the treatment of delinquencies and defaults that can vary among the data sources. As highlighted by Bricker et al. (2015), the G.19 follows the reporting conventions of each institution type, while the CCP contains charged-off loans. Although the SCF should capture delinquencies and defaults, it is possible that respondents are less likely to report loans on which they have not been making payments for a while (Bricker et al., 2015).

Additionally, the SCF differs in the data collection and reporting concept of revolving debt. Whereas the G.19, CCP, and Y-14M estimates collect and report the total revolving debt outstanding (which includes convenience credit), the SCF collects the two components of revolving credit—the revolving component and the convenience credit—separately (Henriques and Hsu, 2013).¹³

Finally, estimates of revolving credit can also be obtained by examining FR Y-14M data.¹⁴ These estimates represent approximately 69 percent of the revolving credit universe reflected in the G.19. The difference is due to the fact that FR Y-14M is submitted only for bank holding companies, savings and loan holding companies, and intermediate holding companies with $100 billion or more in total consolidated assets.¹⁵ Moreover, the FR Y-14M balances we used for comparison do not include charge cards, while the G.19 estimates do. In addition, the Y-14M balances do not include amounts that were charged off.¹⁶

Conclusion

In this note, we discussed several data sources providing the estimation of the revolving credit universe in the U.S. Two of these data sources—G.19 and H.8—publish both levels and growth rates for such credit. Other data sources, such as the CCP, SCF, or FR Y-14M, allow for the construction of level estimates for revolving credit. Although all the discussed series differ on a variety of dimensions—such as representation, frequency, seasonal adjustment, or timing of data collection, among several others—of note, is the similarity in the overall trends exhibited by the estimates based on the different sources discussed, as well as that in the growth rates reported by the G.19 and H.8.

References

Bricker, J., M. Brown, S. Hannon, and K. Pence (2015). "How Much Student Debt is Out There?" FEDS Notes. Washington: Board of Governors of the Federal Reserve System, August 7.

Henriques, A. and J. Hsu (2013). "Analysis of Wealth Using Micro and Macro Data: A Comparison of the Survey of Consumer Finances and Flow of Funds Accounts" FEDS Working Paper No. 2013-46. Washington: Board of Governors of the Federal Reserve System, May.

Lee, D. and W. van der Klaauw (2010). "An Introduction to the FRBNY Consumer Credit Panel". Staff Report no. 479, New York: Federal Reserve Bank of New York, November 10.