Measuring Spending in the Survey of Consumer Finances

October 09, 2026

Measuring Spending in the Survey of Consumer Finances

Aditya Aladangady, Jesse Bricker, Andrew C. Chang, Sarena Goodman, and Gina Li, with assistance from Payton Crawford and Eve Devens¹

The United States has long lacked comprehensive household spending data across all types of households. Data sources whose primary purpose is to measure spending routinely miss the highest-income and wealthiest households (Bee, Meyer, and Sullivan, 2015; Sabelhaus, Johnson, Ash, Swanson, Garner, Greenlees, and Henderson, 2015; Pfeffer, Schoeni, Kennickell, and Andreski, 2016), while data sources that cover the full distribution of income and wealth lack comprehensive spending measures (Attanasio and Pistaferri, 2016; Fisher, Johnson, Smeeding, and Thompson, 2022). This data gap makes it difficult to characterize the distribution of economic well-being fully (National Academies of Sciences, Engineering, and Medicine, 2024), as income and wealth capture only a family's capacity to spend, while spending better reflects their material well-being. As a consequence of this data gap, fundamental questions about spending inequality remain unresolved. In particular, claims of a "K-shaped" post-pandemic recovery-where rich families drive growth as poor families fall behind-are difficult to evaluate because of this data gap.

To fill this gap, we, along with other members of the Federal Reserve staff, introduced new spending questions in the 2025 Survey of Consumer Finances (SCF).² The SCF has long been the leading resource for analyzing household wealth and income in the United States, due to its detailed balance sheet information and oversample of wealthy households. However, it has historically captured only selective outlays-food expenditures, major housing costs, vehicle purchases, and routine debt service-leaving researchers seeking broad spending measures to rely on surveys like the Consumer Expenditure Survey (CE) or the Panel Study of Income Dynamics (PSID), which tend to miss the highest-wealth households, or on administrative data, which are often not drawn from a representative sample frame. The 2025 SCF removes this tradeoff by adding direct measures of total annual spending and vacation spending, allowing researchers to observe families across income, wealth, and spending distributions within a single, representative dataset for the first time.³

We begin by benchmarking the SCF data against the CE and PSID. The CE and PSID collect many categories of spending-especially the CE because of its primary purpose in constructing official price level statistics-and add the categories of spending up to a bottom-up measure of total spending. The SCF, in contrast, collects a few categories of spending plus a top-down total spending amount.

Despite the methodological differences, the SCF's top-down implied aggregate spending figure for 2025 is only slightly below the bottom-up total for 2024 implied by the CE. Moreover, core SCF spending components (food, rent, and new vehicles) track the CE closely over time.⁴ In addition, we find that distributional spending and saving patterns in the SCF roughly align with both the CE and PSID in the bottom 95 percent of the income distribution, which all three sources cover well (Sabelhaus, Johnson, Ash, Swanson, Garner, Greenlees, and Henderson, 2015; Pfeffer, Schoeni, Kennickell, and Andreski, 2016).⁵

Importantly, our benchmarking helps validate that the SCF fills a critical measurement gap in spending for high-income families, allowing us to describe spending patterns over the full economic continuum. We show that the relationship between spending and income extends nearly linearly across the highest income families. These high-income families have an average level of spending that far exceeds both the CE and PSID peaks and surpasses what even an extrapolation of those data sources would project.

Empirical work on economic inequality relies primarily on income or wealth distributions (e.g., Bricker, Henriques, Krimmel, and Sabelhaus, 2016; Smith, Zidar, and Zwick, 2023), leaving the complete distribution of material well-being unobserved.⁶ Here, we provide the first statistics that fully characterize inequality in spending and that examine it alongside income and wealth within a single, unified sample. We find that spending is concentrated, though considerably less so than income or wealth: the top 1 percent of spenders account for 9 percent of total spending, compared with top percent shares of 19 percent for income and 32 percent for wealth. Spending shares rise nonlinearly with household resources, such that the top 10 percent of families ranked by either wealth or income account for a disproportionate 28 percent of total spending. Further, spending composition follows predictable patterns based on expenditure type, with discretionary goods becoming a larger share of spending as income rises. Finally, joint analysis of economic inequality across all three dimensions reveals that 30 percent of the wealthiest families also rank in the top 1 percent of income and spending-roughly half the overlap observed between wealth and income alone. Therefore, a meaningful share of top spenders fall outside the top income and wealth brackets.

Finally, while a single cross-section cannot establish trends, the 2025 SCF also introduced new qualitative questions on changes in income and spending that we use to evaluate whether the 2025 SCF data are consistent with a K-shaped post-pandemic recovery. On net, the data suggest that the position of lower-income families deteriorated a bit relative to higher-income families. While lower-income families were just as likely as higher-income ones to report that their spending was higher than the previous year and that they were better off financially, the share of lower-income families that reported spending decreases, being financially worse off, and lower income during the prior year rose relative to higher-income families. Therefore, the data are only in part consistent with the K-shaped recovery-higher-income families do not appear to be pulling away but at least some lower-income families appear worse off. Future waves of the SCF will allow these dynamics to be measured more directly.

Insights from the new spending measures

The new SCF total spending question asks respondents about all of their family's spending during the last 12 months, including durable and nondurable purchases, but excluding rent, mortgage payments, and retirement contributions.⁷ This scope is conceptually similar to the ex-shelter and retirement contributions measure in the CE. However, the CE collects data on hundreds of different spending areas and aggregates spending from these components, versus the top-down approach in the SCF spending question.⁸ Despite this difference in approach, aggregate spending in the 2025 SCF was $6.58 trillion, just below the comparable aggregate of $7.02 trillion in the 2024 CE.⁹

In separate questions, the SCF collects spending in particular categories: food at home, food away from home, and, new for the 2025 survey, vacations or entertainment. Other SCF questions can also be used to back out spending on rent and new car purchases. Figure 1 confirms that the aggregates for these components, where the definitions align, track the CE over history.¹⁰

As a next step, we compare distributions of spending over income across micro data sources. Here, we focus on both the CE and PSID, two surveys that both measure spending and income, and have good coverage over the bottom 95 percent of the income and wealth distributions (Sabelhaus, Johnson, Ash, Swanson, Garner, Greenlees, and Henderson, 2015; Pfeffer, Schoeni, Kennickell, and Andreski, 2016).

Figure 2 plots average spending by income for the three surveys.¹¹ Spending increases steadily with income across all three surveys with considerable alignment in levels. However, the slope is flatter-and spending in lower income segments is higher-in the CE and PSID, with differences most pronounced in the CE. This pattern is consistent with both (a) potential underreporting of income by low income families in the CE (Meyer and Sullivan, 2003), as the PSID and SCF both measure income relatively well-and (b) potentially larger spending reported in the CE due to the granularity of its spending questions.¹² The PSID spending questions are more granular than the SCF and less granular than the CE, and average spending in the SCF tracks the PSID rather closely.

The key advantage of the SCF relative to other surveys is that it includes an oversample of wealthy households, providing unique insight into spending at the top of the income and wealth distributions. In Figure 2, for example, the SCF is the only dataset that provides coverage of spending at the top of the income distribution (at the far right of the horizontal axis). Average spending for this highest income bin is approximately a linear extension of the relationship seen in the SCF at lower income levels. Average spending for this top-most SCF group is larger than the top segments in both the CE and the PSID, and it exceeds what a linear extrapolation of their distributions would imply. This pattern foreshadows our distributional findings that spending is less concentrated than income and wealth but that aggregate consumption is much more heavily driven by richer families than existing expenditure surveys indicate.

As a final benchmarking exercise, Figure 3 compares saving rates-the portion of after-tax income that remains after accounting for total outlays-across the three datasets over income.¹³ Implied saving rates in the SCF are more comparable to those in the PSID at lower income levels and more comparable to the CE at higher income levels. The higher quality income measure in the SCF may raise saving rates relative to the CE for lower income families, though, as noted above, these families may also under-report spending in the SCF. Saving generally increases with income with saving rates of 50 percent of income or higher for the highest income families across all datasets.¹⁴ These saving rates also imply that roughly one-quarter of families are at their budget constraint-that is, their calculated saving rates are less than or equal to 0.¹⁵

These benchmarking exercises demonstrate that both SCF spending aggregates and the relationships between spending, saving, and income align with external spending data. Furthermore, the SCF captures the full distribution of families by income. We build on this latter point and turn next to a distributional analysis of spending in the SCF.

Spending is right-skewed over both income and wealth and just slightly more evenly spread across wealth than across income (Figure 4, panel A).¹⁶ The highest income and highest wealth families both account for around 28 percent of aggregate spending. Concentration of spending toward the top of the income distribution is much higher in the SCF than in the CE-where the highest decile accounts for a still-high 21 percent of spending (Figure 4, panel B). Differences in population coverage seem to be the driver: in the SCF, two-thirds of total spending in the top income decile was by the top 5 percent by income. Omitting the SCF's wealthy oversample leads to population coverage similar to the PSID and CE (Pfeffer, Schoeni, Kennickell, and Andreski, 2016), and the SCF-ex-wealthy-oversample spending concentration is closer to the CE (dashed blue bars in Figure 4, panel B).¹⁷

Beyond differences in spending levels, how families spend also varies significantly across the income distribution. While we do not observe the full set of goods and services they spend on, some clear patterns emerge from the components we do observe.

The share of spending allocated to food at home is much higher at the bottom of the distribution: around 30 percent of total spending in the lowest two deciles (Figure 5, panel A, light red bars). While actual spending on food at home rises with income (panel B, light red bars), the increase is muted, consistent with food at home being a necessity. As a share of a family's budget, spending on used cars rises to a peak in the 7^th decile of income, then falls (panel A, dark red bars).

By comparison, spending on discretionary components-that is, food away from home, new cars, and vacations-rises as a share of total spending over the distribution (Figure 5, panel A, blue bars), with overall spending in these categories much more concentrated among high-income households (panel B, blue bars). Families in the top decile of income account for more than 30 percent of new car purchases and 40 percent of vacation and entertainment spending (darkest blue bars, panel B). Broadly, the pattern is consistent with non-homothetic spending preferences, where the mix of spending changes with income and wealth (Gaillard, Hellwig, Wangner, and Werquin, 2023), and suggests that relying on just food spending as a proxy for total spending significantly understates distributional variation.

Figure 6 exploits the SCF's unique ability to capture income, wealth, and spending within the same sample. Relative to one another, overall spending is the least concentrated, and wealth is the most, consistent with past literature using imputed SCF expenditures (Fisher, Johnson, Smeeding, and Thompson, 2022) and the PSID (Gaillard, Hellwig, Wangner, and Werquin, 2023).¹⁸ The top 1 percent of spenders account for 9 percent of all spending, while the top 1 percent of earners account for 19 percent of all income, and the richest 1 percent account for 32 percent of all wealth.¹⁹ Spending concentration on discretionary components-vacations, new vehicles, and food away from home-is nearly as concentrated as income, though.

The degree of overlap among income, wealth, and spending in the SCF offers new insight into the joint concentration of economic resources (Fisher, Johnson, Smeeding, and Thompson, 2022) and furthers a long-held measurement goal of integrated wealth, income, and consumption data (National Academies of Sciences, Engineering, and Medicine, 2024). As shown in Table 1, in 2025 57 percent of families in the wealthiest 1 percent were also the top percent of earners, while 46 percent were jointly in the top wealth and spending group. Ultimately, 31 percent of the wealthiest families ranked in the top 1 percent across all three dimensions simultaneously. Broadening to the top 5 percent reveals even tighter alignment: 62 percent of the wealthiest 5 percent of families are the top 5 percent of spenders, with 44 percent of that group aligned on all three metrics. At both levels, the overlap between wealth and spending is somewhat larger than the overlap between income and spending, damping the influence of temporary income shocks. Households in the top percent of income and wealth overlap substantially with each other, but top spender households are more broadly distributed across the population. Consequently, a meaningful share of top spenders falls outside the top income and wealth brackets.

Table 1. Joint distribution of wealth, income, spending

Qualitative measures of spending growth: Evaluating the K-shaped economy narrative

Thus far, we have used the SCF to describe the cross section of spending across both income and wealth distributions, showing that aggregate spending is more concentrated at the top than previously understood. But which income groups are driving the change in spending in the post-pandemic period? Economic growth in the aftermath of the COVID-19 recession has been described as "K-shaped"-where the income and consumption growth of well-off families has diverged from those of other families-even as the evidence of such growth is often based on incomplete data and yields mixed conclusions.²⁰ We cannot construct a change in aggregate spending by income groups, as the 2025 SCF is the first year that collected total spending. However, the 2025 survey also added questions on self-reported qualitative changes in spending, income, and financial well-being, which we use to estimate dynamics in spending and financial well-being by income and wealth.²¹

Focusing first on families' impressions of how their spending changed over the past year, a K-shaped economy predicts that the share of families whose spending increased would rise with income and wealth. However, in the SCF the share reporting an increase in spending is about the same for low-income families as it is for high-income families, as is the case for low-wealth versus high-wealth families (Figure 7).

We can only get a hint at the magnitudes, but the share reporting they "substantially" increased spending also points against spending growth being concentrated at the top deciles (Figure 8). This qualitative evidence is certainly not consistent with spending among high-income or wealth-families increasing disproportionately relative to the rest of families.²² That said, responses to this question do indicate some divergence between low and high income families: the share reporting a decrease in spending is concentrated among lower-income families (with those reporting their spending was unchanged making up the difference: Figure 8).

This pattern also emerges in the remaining qualitative questions about changes in income and financial well-being. Lower-income families are substantially more likely than higher-income families to report that they are financially worse off than a year ago or have lower income than a year ago (Figure 9, panel A). At the same time, a significant amount of families in all income deciles reported increases in spending and well-being, such that the share reporting increases is consistently higher than the share reporting decreases (comparing panel A with panel B).²³ However, unlike the share reporting decreases, the share reporting increases only marginally rises with income.

Overall, the pattern that emerges in Figures 7 to 9 is only in part consistent with a K-shaped recovery. Some lower income families are falling further behind higher income families, though not because of disproportionate gains for higher-income families, but because at least some lower-income families are doing worse.

References

Aguiar, Mark, and Mark Bils (2015). "Has Consumption Inequality Mirrored Income Inequality?" American Economic Review vol. 105, no. 9: 2725-56. DOI: 10.1257/aer.20120599.

Aladangady, Aditya, Jesse Bricker, Andrew C. Chang, Sarena Goodman, Gina Li, Kevin B. Moore, Sarah Reber, Alice Henriques Volz, and Richard A. Windle (2026). Changes in U.S. Family Finances from 2022 to 2025: Evidence from the Survey of Consumer Finances. Washington: Board of Governors of the Federal Reserve System, October, https://doi.org/10.17016/8799.1.

Aladangady, Aditya, Jesse Bricker, Andrew C. Chang, Sarena Goodman, and Gina Li (2026a). "Heterogeneity in the Marginal Propensity to Consume Among U.S. Households," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, October 9.

Aladangady, Aditya, Jesse Bricker, Andrew C. Chang, Sarena Goodman, and Gina Li (2026b). "Informal Support Networks in the Survey of Consumer Finances," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, October 9.

Andreski, Patricia, Geng Li, Mehmet Zahid Samancioglu, and Robert Schoeni (2014). "Estimates of Annual Consumption Expenditures and Its Major Components in the PSID in Comparison to the CE." American Economic Review vol. 104, no. 5: 132-35. DOI: 10.1257/aer.104.5.132.

Attanasio, Orazio P., and Luigi Pistaferri (2016). "Consumption Inequality." Journal of Economic Perspectives vol. 30, no. 2: 3-28.

Bee, Adam, Bruce D. Meyer, and James X. Sullivan (2015). "The Validity of Consumption Data: Are the Consumer Expenditure Interview and Diary Surveys Informative?" In "Improving the Measurement of Consumer Expenditures." Eds: Christopher Carroll, Thomas Crossley, John Sabehaus. University of Chicago Press: Chicago.

Board of Governors of the Federal Reserve System. Division of Research and Statistics, Microeconomic Surveys (2026). "Survey of Consumer Finances.".

Bricker, Jesse, Alice Henriques, Jacob Krimmel, and John Sabelhaus (2016). "Measuring Income and Wealth at the Top Using Administrative and Survey Data." Brookings Papers on Economic Activity no. 1: 261-321.

Browning, Martin, Thomas F. Crossley, and Guglielmo Weber (2003). "Asking Consumption Questions in General Purpose Surveys." Economic Journal vol. 113, no. 491: F540-F567. https://doi.org/10.1046/j.0013-0133.2003.00168.x.

Bureau of Economic Analysis (2025). "Personal Consumption Expenditures by Major Type of Product and Decile." March 14.

Chakrabarti, Rajashri, Thu Pham, Beck Pierce, and Maxim L. Pinkovsky (2026). "Tracking the K-Shaped Economy: Who's Driving Spending?" Liberty Street Economics, Federal Reserve Bank of New York, May 1, 2026.

Dynan, Karen E., Jonathan Skinner, and Stephen P. Zeldes (2004). "Do the Rich Save More?" Journal of Political Economy vol. 112, no. 2: 397-444.

Fisher, Jonathan D., David S. Johnson, Timothy M. Smeeding, and Jeffrey P. Thompson (2022). "Inequality in 3‐D: Income, Consumption, and Wealth." Review of Income and Wealth vol. 68, no. 1: 16-42.

Gaillard, Alexandre, and Christian Hellwig, and Philipp Wangner, and Nicolas Werquin (2023). "Consumption, Wealth, and Income Inequality: A Tail of the Tails (PDF).". Chicago Fed WP 2023-43.

Gindelsky, Marina, and Robert Martin (2025). "The Polarization of Personal Saving." Review of Income and Wealth vol. 71, no. 1: e70004. https://doi.org/10.1111/roiw.70004.

Hacıoğlu Hoke, Sinem, Leo Feler, and Jack Chylak (2024). "A Better Way of Understanding the US Consumer: Decomposing Retail Sales by Household Income," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, October 11.

Hagler, Rees, and Dhiren Patki (2025). "Why Has Consumer Spending Remained So Resilient? Evidence from Credit Card Data." Current Policy Perspectives No. 2025-10 (Federal Reserve Bank of Boston).

Hoham, Benjamin, and Fang Yang (2025). "Consumption Concentration May be up, Adding Slightly to Economic Fragility." Federal Reserve Bank of Dallas.

Horwich, Jeff (2026). "Have US Consumers Gone "K-shaped"? A Review of the Data." Federal Reserve Bank of Minneapolis.

Llanes, Elizabeth, Jeffrey Thompson, and Alice Henriques Volz (2025). "Do the Rich Really Save More? Answering an Old Question Using the SCF with Direct Measures of Lifetime Earnings and an Expanded Wealth Concept." Finance and Economics Discussion Series 2025-097. Washington: Board of Governors of the Federal Reserve System.

Marte, Jonelle (2025, September 16). "Top 10% of Earners Drive a Growing Share of US Consumer Spending." Bloomberg, Accessed August 14, 2026. https://www.bloomberg.com/news/articles/2025-09-16/top-10-of-earners-drive-a-growing-share-of-us-consumer-spending.

Meyer, Bruce D., and James X. Sullivan (2003). "Measuring the Well-Being of the Poor Using Income and Consumption." Journal of Human Resources vol. 38 (2003): 1180-220.

Meyer, Bruce D., and James X. Sullivan (2023). "Consumption and Income Inequality in the United States since the 1960s." Journal of Political Economy vol. 131, no. 2: 247-284.

National Academies of Sciences, Engineering, and Medicine (2024). Creating an Integrated System of Data and Statistics on Household Income, Consumption, and Wealth: Time to Build. Washington, DC: The National Academies Press.

Pfeffer, Fabian T., Robert F. Schoeni, Arthur Kennickell, and Patricia Andreski (2016). "Measuring Wealth and Wealth Inequality: Comparing Two US Surveys." Journal of Economic and Social Measurement vol. 41, no. 2: 103-120.

Sabelhaus, John, David Johnson, Stephen Ash, David Swanson, Thesia Garner, John Greenlees, and Steve Henderson (2015). "Is the Consumer Expenditure Survey Representative by Income?" In "Improving the Measurement of Consumer Expenditures." Eds: Christopher Carroll, Thomas Crossley, John Sabehaus. University of Chicago Press: Chicago.

Smith, Matthew, Owen Zidar, Eric Zwick (2023). "Top Wealth in America: New Estimates Under Heterogeneous Returns." Quarterly Journal of Economics vol.138, no. 1: 515-73. https://doi.org/10.1093/qje/qjac033.

University of Michigan. Survey Research Center, Institute for Social Research (2026). "Panel Study of Income Dynamics. Public Use Dataset." Ann Arbor, MI.

U.S. Bureau of Labor Statistics (2026). "Consumer Expenditure Surveys.".