Consumer Spending Inequality: What Fed Research Reveals About Post-Pandemic Income Gaps

Why Consumer Spending Inequality Matters Now

Consumer spending inequality has become one of the most consequential yet poorly understood dynamics shaping the post-pandemic economy. When economists report that consumer spending remains robust, they are often presenting an aggregate picture that conceals profound differences between income groups. A landmark Federal Reserve working paper (FEDS 2025-050) now provides definitive evidence that this aggregate view has been systematically misleading policymakers, researchers, and the public.

The paper, authored by Sinem Hacıoğlu-Hoke of the Federal Reserve Board, along with Leo Feler and Jack Chylak of Numerator, demonstrates that consumer spending inequality widened dramatically after mid-2021. High-income households continued increasing their spending while low- and middle-income households experienced stagnation or outright decline. Perhaps most troublingly, the methodology used by many leading economic studies completely obscured this divergence. Understanding how economic indicators shape financial decisions has never been more important for individuals navigating this uneven landscape.

This research has profound implications for fiscal policy, monetary policy, and anyone trying to understand the true state of the American consumer. The disconnect between headline economic numbers and lived financial reality for millions of households is not merely an academic curiosity — it represents a fundamental failure of economic measurement that affects real policy decisions.

Inside the Federal Reserve Research Paper

The Federal Reserve study draws on an exceptionally rich dataset: a rolling panel of 150,000 representative U.S. households tracked by Numerator through receipt scanning, digital receipt scraping, and linked loyalty account data. Unlike traditional economic surveys that rely on recall or sampling a small number of households, this dataset captures actual retail transactions from January 2018 through December 2024.

The researchers validated their spending measure against the Census Bureau’s Monthly Advance Retail Sales (MARTS) report, achieving a correlation of 0.94 in spending levels and 0.83 in monthly percent changes. This strong correlation confirms that the micro-level data accurately reflects broader retail spending trends while offering the granularity that aggregate measures cannot provide.

A critical feature of the dataset is self-reported household income in 16 groupings, from less than $20,000 to more than $250,000 annually. Panelists are resurveyed on income approximately every 12 months, with more frequent updates triggered by life events. This allows the researchers to classify households by their actual income rather than relying on geographic proxies — a distinction that turns out to be decisive for understanding consumer spending inequality in the post-pandemic era.

The study’s methodology is rigorous and includes multiple robustness checks. The researchers cross-validated their findings using education levels as an alternative income proxy, conducted Monte Carlo simulations with 1,000 iterations to test sensitivity to income mobility, and performed vector autoregression (VAR) analysis to examine how different income groups respond to economic uncertainty shocks.

The Post-Pandemic Consumer Spending Divergence

The headline finding is striking in its clarity. When households are grouped by self-reported income, a dramatic divergence in consumer spending inequality emerges beginning in mid-2021. Before the pandemic, spending growth was broadly similar across income groups. During the initial pandemic shock in 2020, all groups experienced declines followed by a recovery. The March 2021 stimulus package provided the largest percentage boost to low-income households, consistent with prior research on fiscal multipliers.

However, from mid-2021 onward, the trajectories diverged sharply. By December 2024, relative to January 2018 baseline levels, high-income households (earning more than $100,000 annually) had increased spending by 20.9 percent. Middle-income households ($60,000 to $100,000) had grown spending by 17.0 percent. Low-income households (under $60,000) had increased spending by only 13.5 percent — a gap of 7.4 percentage points compared to their high-income counterparts.

This divergence is not a temporary blip. The spending gap between income groups has persisted for more than three years, suggesting structural rather than cyclical forces are at work. The implications for financial planning and household resilience are substantial, particularly for families whose real purchasing power has been eroding.

Why Zip-Code Income Proxies Fail

The most methodologically important contribution of this consumer spending inequality research involves the failure of zip-code income proxies. Many influential economic studies — including the widely cited Chetty et al. (2023) Economic Tracker — classify households by the median income of their zip code using American Community Survey (ACS) data, rather than by actual household income. This is done because zip-code income data is freely available while household-level income data requires proprietary datasets.

The Fed researchers found that this seemingly reasonable approximation produces catastrophic misclassification rates. In low-income zip codes, only 58.7 percent of residents are actually low-income by self-reported income. A full 26.7 percent are middle-income and 14.6 percent are high-income. In high-income zip codes, the picture is equally distorted: only 32.2 percent of residents are genuinely high-income, while 36.6 percent are low-income and 31.1 percent are middle-income.

Overall, the discrepancy between zip-code aggregated income and self-reported household income ranges from 35 to 75 percent. This misclassification completely masks the post-2021 consumer spending divergence. When the researchers applied zip-code income groupings to the same data, the spending gap between income groups vanished — all groups appeared to have similar spending trajectories, painting a fundamentally misleading picture of consumer financial health.

The Bureau of Labor Statistics Consumer Expenditure Survey provides some household-level data, but the Fed study demonstrates that even established official sources may not capture the granularity needed to identify consumer spending inequality in real time. The misclassification problem is not new; it existed even in 2018 when the ACS data was most current. But it has worsened significantly since the pandemic.

Consumer Spending Inequality by the Numbers

The quantitative findings on consumer spending inequality are worth examining in detail because they reveal just how much economic reality diverges from what standard measurements suggest. The researchers document that income composition within zip codes shifted dramatically between 2018 and 2024.

In 2018, 71.7 percent of low-income households lived in zip codes classified as low-income. By 2024, that figure had dropped to just 58.7 percent. Conversely, the share of high-income households living in low-income zip codes rose from 8.5 percent in 2018 to 14.6 percent in 2024. The biggest shift occurred in 2022, when broad-based wage increases pushed many households into higher income brackets without corresponding changes in residential location.

The misclassification rates tell a damning story for economic research methodology. For low-income groups, 41.3 percent of households are misclassified when using zip-code proxies. For middle-income groups, the misclassification rate reaches 69.2 percent. For high-income groups, it stands at 67.8 percent. These are not marginal errors — they represent a fundamental breakdown in how economists categorize and analyze consumer behavior by income.

The VAR analysis adds another dimension. When hit by an economic uncertainty shock, low-income household spending (measured by self-reported income) declines by approximately 2.5 percentage points at trough — sensible because these households spend primarily on essentials and have limited capacity to cut back. Using zip-code income proxies, however, low-income households appear to cut back almost as much as high-income households, roughly 10 percentage points. This implausible result could lead to severely miscalibrated policy responses.

How Income Misclassification Distorts Policy

The policy consequences of consumer spending inequality being hidden by measurement error are significant. When the Federal Reserve sets monetary policy, it relies on assessments of consumer spending strength and labor market conditions. If aggregate measures suggest that spending is broadly strong across income groups, there is less urgency to consider the distributional effects of interest rate decisions.

The reality exposed by this research is that high-income households have been disproportionately driving spending resilience. These households benefited from rising home values, investment gains, and higher interest income during the rate-hiking period. Meanwhile, low- and middle-income households faced the headwinds of savings depletion, expiration of government support programs, and the compounding effects of inflation on essential goods.

Fiscal policy is similarly affected. The March 2021 stimulus demonstrated that targeted transfers to low-income households produce large spending effects — the researchers confirm that this was the period of greatest spending growth for the lowest income group. But without accurate real-time data on consumer spending inequality, policymakers cannot assess when and how much additional support might be needed, or whether existing programs are reaching the households experiencing the greatest financial stress.

For anyone following economic policy debates, understanding how data quality shapes decisions is essential. The tools we use to measure the economy determine the policies we design, and this research reveals a systematic blind spot that has persisted for years. Exploring how government economic reports influence markets provides important context for investors and analysts alike.

The Role of Savings Depletion and Government Support

Several factors converged to create the persistent consumer spending inequality documented in this study. The first and arguably most important is the depletion of pandemic-era excess savings. During 2020 and early 2021, government transfer payments — including stimulus checks, enhanced unemployment insurance, and expanded child tax credits — combined with reduced spending opportunities to create a savings buffer across income groups.

However, these excess savings were not distributed equally, and they were depleted at vastly different rates. Research cited in the Fed paper, including work by Abdelrahman et al. (2024), shows that low-income households exhausted their pandemic savings far more quickly than high-income households. By mid-2021, many low-income families had already spent through their buffer and were facing a return to pre-pandemic financial constraints — or worse, given rising prices.

The expiration of government support programs compounded the challenge. SNAP emergency allotments, which had provided additional food assistance during the pandemic, were phased out. Enhanced unemployment insurance benefits ended in September 2021. The expanded child tax credit, which provided monthly payments of $250 to $300 per child, expired at the end of 2021. Each of these program expirations disproportionately affected low- and middle-income households, removing income support just as their savings buffers were running dry.

Meanwhile, high-income households experienced positive wealth effects. Rising home values increased perceived wealth, stock market gains boosted investment portfolios, and higher interest rates — implemented by the Fed to combat inflation — actually increased interest and investment income for households with significant savings and investment assets. This asymmetric impact of monetary tightening further widened consumer spending inequality.

Consumer Spending Inequality and Inflation

The interaction between consumer spending inequality and inflation adds another critical dimension to the analysis. The researchers note that low-income households face higher effective inflation rates than high-income households because they allocate a larger share of their spending to categories that experienced the greatest price increases — food, energy, housing, and transportation.

Research by Cavallo and Kryvtsov (2024), cited in the Fed paper, demonstrates that differential inflation experiences mean the real spending divergence between income groups is likely even wider than the nominal figures suggest. When a low-income household’s spending increases by 13.5 percent in nominal terms but the prices they face have risen by a greater percentage than the overall CPI, their real purchasing power may have actually declined.

This finding has important implications for how we interpret consumer spending data. Headline inflation measures like the Consumer Price Index represent an average across all consumption baskets. But no individual household consumes the average basket. Low-income households spend proportionally more on food at home (where prices surged), rent (which rose sharply in many markets), and gasoline (which experienced significant volatility). High-income households, by contrast, allocate more spending to services, travel, and discretionary categories where price increases were more moderate or where they have greater flexibility to substitute.

The net result is that consumer spending inequality measured in real terms — adjusted for the prices each income group actually faces — is substantially greater than the already concerning nominal figures. This distinction matters enormously for policy design and for understanding the true financial health of American consumers across the income spectrum.

Better Data for Better Economic Decisions

The Federal Reserve researchers make a compelling case for transforming how economic data is collected and used. Their micro-level panel data, covering 150,000 households with actual transaction records and self-reported income, can be updated on a weekly basis. This stands in stark contrast to the American Community Survey data used for zip-code income classification, which covers the 2014-2018 period and is increasingly stale.

The authors advocate for what might be called a data revolution in economic measurement. Rather than relying on geographic aggregation that produces 35 to 75 percent misclassification rates, researchers and policymakers should invest in alternative datasets that capture household-level heterogeneity. The technology to do so already exists — receipt scanning apps, digital transaction data, and linked financial accounts can provide unprecedented granularity on consumer spending behavior.

This is particularly important during periods of economic disruption, when consumer behavior changes rapidly and the composition of income groups shifts due to job losses, wage changes, and government interventions. The post-pandemic period has demonstrated that standard measurement tools are most likely to fail precisely when accurate data matters most.

The robustness of the findings is reinforced by the education-based analysis. When the researchers classified households by educational attainment rather than income — a more stable characteristic over time — they found the same pattern. Households with high school education or less showed spending patterns similar to low-income groups, while households with higher education tracked high-income spending behavior. This cross-validation strengthens confidence that consumer spending inequality is real and not an artifact of measurement choices.

What This Means for Households and Investors

For individual households, this research validates what many have felt intuitively: the economic recovery has been deeply uneven. If you are a low- or middle-income household that has seen your spending power erode while headline numbers proclaim economic strength, the data now confirms that your experience is real and widespread. The aggregate numbers were masking your reality behind the stronger spending of higher-income households.

For investors and financial analysts, the implications are equally significant. Retail spending data, one of the most closely watched economic indicators, may present a misleadingly optimistic picture when high-income spending masks low-income strain. Sectors dependent on low- and middle-income consumers may be more vulnerable than aggregate data suggests, while luxury and premium segments may show greater resilience.

The research also highlights the importance of granular, real-time data for decision-making at every level. Whether you are a policymaker calibrating stimulus programs, a business planning inventory and pricing strategy, or an individual making financial decisions, the quality of the data you rely on determines the quality of the decisions you make. The era of accepting aggregate proxies as good enough should be ending.

Looking ahead, the consumer spending inequality documented in this research shows no signs of narrowing. Without targeted policy interventions, rising asset values will continue to benefit high-income households while low-income households face the ongoing challenges of elevated prices, limited savings, and reduced government support. Understanding this dynamic is essential for anyone seeking to navigate — or improve — the current economic landscape.

For deeper analysis of how economic research translates into actionable financial insights, explore our interactive library of research summaries and financial guides.