## Drivers of Post-COVID Private Consumption in the U.S. — Section 1

## Source details

**Canonical URL:** [Drivers of Post-COVID Private Consumption in the U.S. — Section 1](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024128-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024128-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024128-print-pdf.pdf.json)

---

### Abstract / Overview
- Private consumption recovered swiftly from the pandemic trough and ran above the pre-pandemic trend despite sharp interest rate increases.
- Main drivers identified: excess savings from the pandemic; large increases in household wealth (especially housing); solid real income gains.
- Compared with pre-COVID estimates, the marginal propensity to consume out of housing wealth is substantially higher; large housing price gains made the wealth effect a key driver of post-pandemic consumption growth.
- JEL Classification Numbers: D12; D14; E21; E24; G51; R39
- Keywords: private consumption; marginal propensity to consume; income; saving; wealth; post-COVID recovery.

### Introduction: Context and Key Results
- Private consumption has been the primary engine of U.S. GDP growth since the pandemic; real personal consumption expenditures closely track GDP growth.
- Key empirical findings (state- and household-level):
  - Solid labor income growth supports consumption across income distribution; post-pandemic marginal propensity to consume out of that income is somewhat higher.
  - Housing wealth plays a much more important role in driving consumption growth among better-off homeowners relative to the past.
  - Excess savings accumulated during the pandemic supported consumption well into the recovery.
  - Government transfers supported consumption.
  - Aggregate magnitudes: housing and possibly other types of wealth are the most dominant forces underpinning consumption growth.
- Paper organization: Section 2 stylized facts; Section 3 potential drivers; Section 4 data and empirical strategy; Section 5 state-level results; Section 6 household-level results; Section 7 macro contribution calculations; Section 8 outlook and conclusions.

### Stylized Facts (Section 2)
- Consumption categories:
  - Services: most severe decline during lockdowns; solid recovery back to pre-COVID trend after reopening.
  - Goods: brief drop during COVID; rebound surpassed pre-COVID trend.
  - At highest cyclical levels in 2021Q2:
    - Durable goods consumption exceeded about 20 percent their pre-COVID trend level.
    - Nondurable goods consumption exceeded about 10 percent their pre-COVID trend level.
  - Elevated durable goods consumption likely reflects pent-up demand.
- Subcategory patterns:
  - Recreational durable goods hardly declined and stayed on an elevated trend (substitutes for services).
  - Health care, restaurant, and hotel spending rebounded fast to catch up pent-up services demand.
- By income distribution:
  - Bottom decile: averaged annualized consumption growth of 6 percent between 2019 and 2022 (more than doubling the 2014–2019 rate).
  - Top two quintiles accounted for almost half of the average 5 percent consumption growth between 2019 and 2022 due to larger consumption shares.
  - The U.S. consumer economy is primarily powered by relatively well-off households, increasingly so post-pandemic.

### Potential Drivers (Section 3)
- Labor market:
  - Employment and wage growth raised disposable incomes and job security.
  - Unemployment persistently below 4 percent for an extended period.
  - Workers’ earnings outpacing U.S. inflation in most recent quarters supported consumer confidence and spending.
- Fiscal stimulus and transfers:
  - Stimulus checks and enhanced unemployment benefits amounted to approximately 4-5 percent of GDP per year during 2020-21.
  - Fiscal transfers improved household balance sheets; some households used transfers to bolster savings or repay high-cost debt.
- Excess savings:
  - Pandemic support generated an unprecedented surge in personal savings above pre-pandemic trends.
  - Empirical evidence indicates a drawdown of pandemic-related “excess” savings since mid-2021 of about 2.1 trillion dollars cumulatively.
- Wealth effects:
  - Housing wealth increased by around 50 percent since the pandemic; stock market wealth increased by around 25 percent in aggregate terms.
  - Housing price appreciation bolstered home equity and financial security and enabled access to home equity credit to finance consumption.
  - Stock market appreciation increased perceived financial security and could prompt higher spending.

### Data and Methodology (Section 4)
- State-level analysis:
  - Sample period: 1990Q1 to 2023Q4.
  - Income components from BEA Regional Economic Accounts; housing wealth constructed using home ownership rate × number of households × Freddie Mac House Price Index by state.
  - Variables deflated by the PCE price index.
  - Real state-level consumption proxied by retail employment; elasticity between real consumption and retail employment assumed to be one for interpretation.
  - Baseline state regression: lnc_st = β_1 lny_st + β_2 lnh_st + β_3 lna_st + α_s + q_t + ε_st (c proxied by retail employment; y non-asset income; a asset-based income; h housing wealth).
  - Unit root tests reject nonstationarity (Table A1).
- Household-level analysis:
  - Data source: U.S. Consumer Expenditure Survey (CE), 2013–2022 (≈7,000–8,000 households per year).
  - Consumption: total consumption expenditure (food, housing-related, transportation, healthcare, entertainment, etc.); categories grouped into five groups.
  - Income measures: total income after tax; and income after tax excluding transfers and asset income.
  - Wealth measures: housing value; savings; other financial assets (gross and net), with liabilities excluding mortgage/lease.
  - Values converted to 2017 real dollars using the PCE price index.
  - Baseline household regression: lnC_jt = α_j + β_1 lnY_jt + β_2 lnH_jt + β_3 lnI_jt + β_4 lnA_jt + q_t + ε_jt (cohort pseudo panel fixed effects to mitigate time-invariant heterogeneity).
  - Heterogeneity recognized: consumption responses differ across income and wealth distributions due to liquidity constraints.

### State-Level Results: Baseline and Time-Varying Elasticities
- Pre-COVID baseline (1990Q1–2019Q4):
  - Elasticity of consumption to non-asset income: 0.5 (interpreted as MPC of 50 cents on one dollar).
  - Elasticity of consumption to housing wealth: 0.06 (interpreted as MPC of 3 cents on one dollar; not significant).
- Post-COVID sample (2020Q1–2023Q4) highlights (Table 1):
  - Income (non-asset) coefficient: 0.0047 (0.052) in 20Q1-23Q4 (not significant).
  - Housing wealth coefficient: 0.1463*** (0.019) in 20Q1-23Q4.
  - Column (3) adds cumulative income since COVID, 1-quarter lag: Asset-based income 0.1688** (0.068); Cumulative income since COVID, 1-quarter lag 0.2930** (0.129); Housing wealth 0.1352*** (0.018).
  - Wage in reopening: 0.2731*** (0.074); income other than wage and dividend/interest/rent: -0.0599** (0.023); housing wealth: 0.0887*** (0.018).
  - Monetary policy and supply disruptions: Federal Funds Rate -0.0076*** (0.002); global supply chain pressure index -0.0072*** (0.001).
- Key state-level findings:
  - Restricting to 2020Q1–2023Q4 yields a larger housing wealth elasticity, roughly doubling relative to pre-COVID.
  - Cumulative income since 2020Q1 used as proxy for excess savings shows significant elasticity 0.2930**.
  - Wage elasticity at reopening about 0.3, translating to an MPC of about 40 cents on one dollar wage increase.
  - Income other than wage and dividend/interest/rent sometimes shows negative elasticity, possibly reflecting reverse causality from government transfers.
  - Robustness with imputed real consumption yields similar elasticities.
- Time-varying quarter-by-quarter specification:
  - Income elasticity lower at reopening (2021Q1) then recovers.
  - Housing wealth elasticity roughly at pre-COVID level prior to 2022Q2; since 2022Q2 several quarters show notably higher housing wealth elasticity than pre-COVID and post-2001-recession levels — indicating housing wealth elasticity increased after COVID.

### Household-Level Results: Baseline Elasticities and Heterogeneity
- Baseline elasticities (Equation 2; Table 2):
  - Consumption elasticity w.r.t. income after tax: ≈0.3 across full sample and subperiods (examples: 0.345*** (0.009); 0.339*** (0.012); 0.340*** (0.025); 0.352*** (0.013)).
  - Consumption elasticity w.r.t. home value (housing wealth): ranges from 0.1 to 0.18 and is higher post-COVID (examples: 0.100*** (0.006); 0.153*** (0.018); 0.180*** (0.013)).
  - Savings and stocks elasticities modest and not consistently significant (example: Savings 0.003*** (0.000); Stocks 0.003*** (0.000)).
  - L.Cumulative saving since Covid shows rising importance: 0.004 (0.003) pre-COVID proxy; 0.024** (0.009) COVID; 0.052*** (0.011) post-COVID in specific specs.
  - Dividend/interest/rent elasticities positive and significant (e.g., 0.018*** (0.001); 0.013*** (0.003); 0.012*** (0.002)).
- Heterogeneity by home-ownership status (Equation 3; Table 3):
  - Income elasticities (selected):
    - Homeowner w/ mortgage: 0.392*** (Pre-COVID), 0.450*** (COVID), 0.393*** (Post-COVID).
    - Homeowner w/o mortgage: 0.424*** (Pre-COVID), 0.382*** (COVID), 0.441*** (Post-COVID).
    - Renter: 0.429*** (Pre-COVID), 0.443*** (COVID), 0.443*** (Post-COVID).
    - Occupant w/o rent payment: 0.363*** (Pre-COVID), 0.574*** (COVID), 0.069 (Post-COVID).
  - Housing value elasticities (selected):
    - Homeowner w/ mortgage: 0.166*** (Pre-COVID), 0.176*** (COVID), 0.214*** (Post-COVID).
    - Homeowner w/o mortgage: 0.072*** (Pre-COVID), 0.114*** (COVID), 0.149*** (Post-COVID).
  - Elasticities w.r.t. cumulative implied saving since COVID (selected):
    - Homeowner w/ mortgage: 0.011*** (Pre-COVID), 0.035*** (COVID), 0.056*** (Post-COVID).
    - Homeowner w/o mortgage: 0.006 (Pre-COVID), 0.039*** (COVID), 0.107*** (Post-COVID).
    - Renter: -0.006 (Pre-COVID), 0.042*** (COVID), 0.039*** (Post-COVID).
- Interpretation:
  - Better-off households accumulated more income and housing wealth and saw increases in elasticities post-COVID, implying these groups contributed disproportionately to aggregate consumption growth.

### Spending Categories, MPCs, and Heterogeneous Responses (Section 3 details)
- MPCs from saving by income:
  - Method: multiply estimated elasticities by consumption-to-savings ratio by income quintile.
  - Main finding: saving MPCs highest among lower-income households; low-income households spend ≈40-45 cents for every dollar increase in cumulative savings.
  - Top 40 percent by income: respond ≈20-25 cents for every dollar increase in cumulative savings.
- Expenditure-category patterns (five groups: recreation; automobile purchase; healthcare; home food; others):
  - Higher income elasticities for recreation, automobile, and healthcare; lower for home food.
  - Post-COVID increases in income elasticities are pronounced for recreation.
  - Selected recreation elasticities (Pre-COVID vs Post-COVID where noted):
    - Recreation income elasticities (Post-COVID examples): Homeowner w/ mortgage 0.655*** (Post-COVID); Homeowner w/o mortgage 0.653*** (Post-COVID); Renter 0.762*** (Post-COVID).
    - Home value elasticities for Recreation (Post-COVID examples): Homeowner w/o mortgage 0.328*** (Post-COVID).
    - L.Cumulative implied saving since COVID (Recreation, Post-COVID examples): Homeowner w/ mortgage 0.151*** (Post-COVID); Homeowner w/o mortgage 0.175*** (Post-COVID).
  - Illustrative post-COVID magnitudes: recreational spending among homeowners without mortgage estimated elasticity ≈0.3 w.r.t. housing wealth and ≈0.2 w.r.t. excess saving.

### Robustness Checks
- Including other financial assets (gross/net): does not change housing wealth elasticities; elasticities for other financial assets are notably small.
- Controlling for fixed-rate mortgage and student loans:
  - Households with fixed-rate mortgage consume slightly more; coefficients become insignificant in post-COVID and cumulative saving specifications.
  - No statistically significant relationship between consumption and student loan status in most specs; interpret cautiously due to possible underreporting and nonlinear debt effects.
- Imputed real consumption (state-level) and unit-root tests reported in appendices confirm robustness of main state-level patterns.

### Aggregate Decomposition and Macro Contributions (Section 3 aggregation exercise)
- Method: apply household-level elasticities to macro changes in after-tax wages and salary income, current transfers, dividend/interest/rental income, and excess pandemic saving drawdowns; macro data from BEA and Federal Reserve Board; deflated with PCE price index.
- Key macro magnitudes (percent deviation from 2019Q4):
  - 2020Q2 trough: consumption 10 percent below 2019Q4; loss in real incomes explains only a small share; residual dominated by mobility restrictions and supply shocks.
  - By end-2022: aggregate consumption ≈7 percent above pre-pandemic level, decomposed as:
    - 4.6 percent explained by rise in housing wealth (using post-COVID household estimates).
    - 1.6 percent explained by higher real household incomes.
    - 1.1 percent explained by drawdown of excess pandemic savings.
  - By end-2023 (additional increase relative to 2022):
    - Higher real incomes contributed additional 1.1 percent (computed 2.7 − 1.6 = 1.1percent).
    - Housing wealth contributed additional 0.6 percent (computed 5.2 − 4.6 = 0.6 percent).
    - Excess saving drawdown contributed 0.5 percent relative to end-2019.
  - One fifth of consumption growth in 2023Q4 relative to 2019Q4 remains unexplained (grey residual).
- Cumulative draw-down of excess saving implied ≈2 Trn Dollar (nominal), close to the 2.1 Trn Dollar estimate referenced.
- External validity:
  - Residuals largest and negative during lockdown; close to zero when consumption recovered; positive but small in 2023Q4 when consumption unusually strong.
  - Micro estimates applied out-of-sample (2023) explain around 80 percent of observed aggregate consumption growth in 2023.

### Main Takeaways and Outlook (Policy-Relevant Implications)
- Core conclusion: housing wealth has become the most important driver of private consumption in recent years; consistent across state- and household-level analyses.
- Distributional pattern: consumption growth concentrated among upper-income households.
- Hypotheses for larger post-pandemic propensity to consume out of housing wealth (for future research):
  - Housing wealth more widely held across segments (including increased home ownership among young), raising aggregate consumption sensitivity.
  - Estimated housing wealth effect may partially capture under-reported non-housing financial wealth correlated with housing wealth (SCF shows homeowners hold much larger non-housing financial wealth).
  - Cryptocurrency wealth correlated with housing wealth and associated with larger MPCs in related work.
- Policy-relevant implications:
  - Continued disinflation, resilient labor market, and robust wage growth support consumption via higher real incomes.
  - Continued housing price appreciation (as interest rates decline and demand rises) could provide further tailwind to consumption and growth since households are more responsive to wealth changes.
  - Conversely, repricing of housing and other financial assets poses downside risks — a sharp wealth retrenchment could lead to notable consumption weakness.

*IMF Working Paper WP/24/128 — Western Hemisphere Department, June 2024.*

### Section 1

### Drivers of Post-COVID Private Consumption in the U.S. — Section 1

### Abstract / Overview
- Private consumption in the U.S. has recovered swiftly from the pandemic trough and has been running above the pre-pandemic trend even as interest rates rose sharply.
- Using both state- and household-level data, the paper finds that excess savings from the pandemic, large increases in household wealth (especially housing), along with solid real income gains contributed to strengthening post-pandemic consumption.
- Compared with pre-COVID estimates, the marginal propensity to consume out of housing wealth is substantially higher, which, together with large gains in housing prices, made the wealth effect a key driver for post-pandemic consumption growth.
- JEL Classification Numbers: D12; D14; E21; E24; G51; R39
- Keywords: private consumption; marginal propensity to consume; income; saving; wealth; post-COVID recovery.

### Introduction: Context and Key Results
- The U.S. economy’s strong GDP growth since the pandemic has been primarily driven by private consumption; the growth in real personal consumption expenditures closely tracks GDP growth (Figure 1).
- Understanding U.S. GDP growth requires understanding U.S. consumption growth, given the U.S. is a large, relatively closed economy where private consumption has been the main engine of economic growth in the post-war period.
- Key empirical findings established using state- and household-level data:
  - Solid labor income growth has been supporting consumption across all segments of the income distribution, with the post-pandemic period registering a somewhat higher marginal propensity to consume out of that income.
  - Relative to the past, housing wealth has been playing a much more important role in driving consumption growth among better-off homeowners.
  - Excess savings built up during the pandemic have supported consumption long into the current juncture.
  - Government transfers played a role in supporting consumption.
  - In terms of overall economic magnitudes, the contribution of housing and possibly other types of wealth has been the most dominant force underpinning aggregate consumption growth.
- Paper organization summary: Section 2 stylized facts; Section 3 overview of potential drivers; Section 4 data and empirical strategy; Section 5 state-level results; Section 6 household-level results; Section 7 macro contribution calculations; Section 8 outlook and conclusions.

### Stylized Facts (Section 2)
- Consumption categories:
  - Services experienced the most severe decline during lockdowns and staged a solid recovery back to the pre-COVID trend after reopening.
  - Goods consumption dropped briefly during COVID but rebounded swiftly and surpassed the pre-COVID trend.
  - At their highest cyclical levels in 2021Q2, durable goods and nondurable goods consumption exceeded about 20 and 10 percent their pre-COVID trend levels, respectively.
  - Elevated durable goods consumption likely reflects postponed durables consumption during lockdown creating pent-up demand.
- Subcategory patterns:
  - Recreational durable goods hardly experienced any setback during lockdown and continued on an elevated trend (substitutes for services).
  - Spending on health care, restaurant, and hotel rebounded fast alongside reopening to catch up pent-up services demand.
- Consumption by income distribution:
  - The bottom decile saw an averaged annualized consumption growth of 6 percent between 2019 and 2022, more than doubling the rate during 2014 - 2019.
  - High-income households increased consumption at a lower rate than the bottom decile, but because of their larger consumption share, the top two quintiles accounted for almost half of the average 5 percent consumption growth between 2019 and 2022.
  - The U.S. consumer economy is primarily powered by the relatively well-off, even more so in recent post-pandemic years.

### Potential Drivers (Section 3)
- Labor market:
  - Robust employment and wage growth raised disposable incomes and job security, supporting consumer spending.
  - Employment rates continued to exceed historical averages despite a slowdown from the 2021 peak; unemployment has been persistently below 4 percent for an extended period.
  - Workers’ earnings outpacing U.S. inflation rates in most recent quarters could further support consumer confidence and spending.
- Fiscal stimulus and transfers:
  - Government interventions including stimulus checks and enhanced unemployment benefits amounted to approximately 4-5 percent of GDP per year during 2020-21.
  - Fiscal transfers improved household balance sheets for many households, with some households using transfers to bolster savings or repay high-cost debt.
- Excess savings:
  - Pandemic fiscal support and income preservation generated an unprecedented surge in personal savings above pre-pandemic trends.
  - Empirical evidence indicates a drawdown of pandemic-related “excess” savings since mid-2021 of about 2.1 trillion dollars cumulatively (Abdelrahman and Oliveira (2023a)).
- Wealth effects:
  - Since the pandemic, housing wealth increased by around 50 percent and stock market wealth increased by around 25 percent in aggregate terms.
  - Housing price appreciation bolstered homeowners’ home equity and increased financial security; higher home values potentially enabled access to home equity loans or lines of credit to finance consumption.
  - Stock market wealth appreciation similarly increased perceived financial security and could prompt higher spending.

### Data and Methodology (Section 4)

- State-level analysis:
  - Dataset period: 1990Q1 to 2023Q4.
  - Income and components sourced from the Regional Economic Accounts of BEA: total income, wage, dividend/interest/rent income (asset-based income), and other income.
  - Housing wealth constructed as: home ownership rate (Housing Vacancies and Home-ownership dataset of Census) × number of households (State and County Intercensal Tables of Census for years prior to 2010 and American Community Survey from 2010 onward) × Freddie Mac House Price Index for each state.
  - Income variables and housing wealth converted to real values by deflating by the personal consumption expenditure (PCE) price index.
  - Real state-level consumption proxied by retail employment (literature precedent: Guren et al. (2021)); elasticity between real consumption and retail employment assumed to be one when interpreting empirical results as consumption responses.
  - Control variables: mobility index during COVID (Google Mobility), Federal Funds Rate, and global supply chain pressure index (Federal Reserve Bank New York).
  - Baseline state-level regression (Equation 1):
    lnc_st = β_1 lny_st + β_2 lnh_st + β_3 lna_st + α_s + q_t + ε_st
    - c is proxied with retail employment.
    - y is income (excluding income from assets: dividend, interest, and rent income).
    - a is asset-based income.
    - h is housing wealth.
    - α_s and q_t represent state fixed effects and quarter fixed effects, respectively.
    - Coefficients of primary interest: β_1 and β_2 (elasticities of consumption to non-asset income and housing wealth).
    - Unit root tests (Table A1) reject nonstationarity.
- Household-level analysis:
  - Data source: U.S. Consumer Expenditure Survey (CE) from 2013 to 2022 (approximately 7,000 to 8,000 households per year).
  - Consumption definition: total consumption expenditure including food (home and away), housing-related expenditure (mortgage/rent, property tax and insurance, utilities, home repairs and furnishing), transportation (personal and public and maintenance), healthcare, entertainment, and others.
  - Expenditure categories further grouped into: (i) recreational expenditure (vacation transportation and accommodation, entertainment, personal care); (ii) automobile purchases; (iii) healthcare expenditure; (iv) home food expenditure; and (v) other expenditures.
  - Income measures: total income after tax; and income after tax excluding transfers and income derived from assets (dividend, interest, rent).
  - Wealth measures: housing value; saving; other financial assets (stocks, mutual funds, bonds; retirement accounts, insurances) examined in gross and net terms (liabilities include credit card debt, student loans, and other loans excluding mortgage and lease).
  - Nominal values converted to 2017 real dollars using the PCE price index.
  - Baseline household-level regression (Equation 2):
    lnC_jt = α_j + β_1 lnY_jt + β_2 lnH_jt + β_3 lnI_jt + β_4 lnA_jt + q_t + ε_jt
    - C is consumption.
    - Y is income after tax (excluding transfers and income from dividend, interest, and rent in some specifications).
    - H is housing wealth.
    - I is a matrix of other income types excluded from Y in some specifications such as transfers and income from dividend, (gross) interest, and rent.
    - A is a matrix of asset variables such as saving and other non-housing assets.
    - α_j denotes cohort fixed effects (cohorts constructed by year of birth, gender, education of the reference person, and home-ownership status to form a pseudo panel).
    - q_t denotes time fixed effects.
  - Cohort-based pseudo panel approach used to mitigate bias from unobserved household time-invariant heterogeneity (Browning, Deaton and Irish (1985); Deaton (1985)).
  - Analysis recognizes consumption responses differ across income and wealth distributions due to liquidity constraints.

*IMF Working Paper WP/24/128 — Western Hemisphere Department, June 2024.*

### Section 2

### wpiea2024128-print-pdf - Section 2

### Methodology: Allowing Heterogeneous Elasticities at Household Level
- Specification used to allow elasticities to vary by home-ownership status:
  - lnC_jt = α_j + β_1^d DlnY_jt + β_2^d DlnH_jt + β_3 lnI_jt + β_4 lnA_jt + q_t + ε_jt (Equation 3)
  - D denotes household groups by home-ownership status.
- β_1 and β_2 measure the elasticity of consumption with respect to income and housing wealth and are the main focus.
- Pseudo panel fixed-effect estimates are used to control for observed and unobserved time-invariant characteristics.
- Note on wealth vs. income: wealth reflects permanent income and demonstrates less volatility than income.

### Complementarity between State- and Household-level Analyses
- State-level data:
  - Timely release provides latest observations on consumption, income, and wealth.
  - Variables are not self-reported and are less prone to measurement errors present in survey data.
- Household-level data:
  - Provides granular information to examine heterogeneity in consumption responses across household groups and budget constraints.
  - Subject to measurement errors: recall bias in expenditures, under-reporting and non-response bias in income, assets, and liabilities.
- Estimated elasticities from both levels represent conditional correlations (contingent upon wealth and transfers), not causal effects.
- Analysis illustrates shifts in these conditional correlations before and after the COVID-19 pandemic.

### State-Level Results: Baseline and Time-Varying Elasticities
- Baseline panel regression (1990Q1–2019Q4; state and quarter fixed effects):
  - Elasticity of consumption for non-asset income: 0.5
  - Elasticity of consumption for housing wealth: 0.06 (not significant)
  - Interpretation in MPC: MPC of 50 cents on one dollar increase in non-asset income and 3 cents on one dollar increase in housing wealth.
- Table 1 highlights (columns summarized):
  - Column (1) sample: 90Q1-19Q4; Income (non-asset) coefficient: 0.5094*** (standard error 0.111); Housing wealth: 0.0588 (0.041); Observations: 6,120; Within R-squared: 0.663.
  - Column (2) sample: 20Q1-23Q4; Income (non-asset) coefficient: 0.0047 (0.052); Housing wealth: 0.1463*** (0.019); Observations: 816; Within R-squared: 0.176.
  - Column (3) adds cumulative income since COVID, 1-quarter lag: Asset-based income coefficient: 0.1688** (0.068); Cumulative income since COVID, 1-quarter lag: 0.2930** (0.129); Housing wealth: 0.1352*** (0.018); Observations: 816; Within R-squared: 0.189.
  - Column (4) separates wage: Wage coefficient: 0.2731*** (0.074); Income other than wage and dividend/interest/rent: -0.0599** (0.023); Housing wealth: 0.0887*** (0.018); Mobility, retail/transit/work: 0.0006*** (0.000); Observations: 816; Within R-squared: 0.270.
  - Column (5) drops quarter FE but includes Federal Funds Rate and supply disruptions: Asset-based income 0.1571** (0.067); Income other than wage and dividend/interest/rent: -0.0326 (0.057); Housing wealth: 0.0861*** (0.017); Mobility: 0.0014*** (0.000); Federal Funds Rate: -0.0076*** (0.002); Global supply chain pressure index: -0.0072*** (0.001); Observations: 816; Within R-squared: 0.654.
- Key state-level findings:
  - Restricting sample to 2020Q1–2023Q4 yields a larger elasticity of housing wealth, roughly doubling relative to pre-COVID.
  - Binscatter (Figure 5) shows slope steepens since COVID; standard errors are small.
  - Cumulative income since 2020Q1 used as a proxy for excess savings; column (3) shows significant elasticity from this savings proxy (0.2930**).
  - Wage elasticity in reopening: about 0.3, translating to an MPC of about 40 cents on one dollar wage increase (column 4).
  - Income other than wage and dividend/interest/rent shows negative elasticity in some specifications, possibly due to reverse causality from government transfers.
  - Monetary policy and supply disruptions: Federal Funds Rate negative coefficient (-0.0076***); supply disruptions negative coefficient (-0.0072***).
- Robustness:
  - Imputed real consumption (using ln(real consumption)_st = γ ln(retail employment)_st + λ T_t + α_s + ε_st (Equation 4)) yields elasticities similar to baseline.
- Time-varying quarter-by-quarter specification:
  - ∆lnc^19Q4_st = γ_1 ∆lny^19Q4_st + γ_2 ∆lnh^19Q4_st + α + ε^19Q4_st, for each quarter t starting from 21Q1 (Equation 5).
  - Figure 6: Income elasticity lower at reopening (2021Q1) then recovers; housing wealth elasticity roughly at pre-COVID level prior to 2022Q2, but since 2022Q2 several quarters show notably higher housing wealth elasticity than pre-COVID and post-2001-recession levels.
  - Indicates housing wealth elasticity increased after COVID.

### Household-Level Results: Baseline Elasticities and Heterogeneity
- Baseline results (Equation 2; Table 2):
  - Across full sample and subperiods (All; Pre-COVID 2013Q1-2019Q4; COVID 2020Q1-2021Q1; Post-COVID 2021Q2-2022Q4):
    - Consumption elasticity with respect to income after tax: ~0.3 (e.g., 0.345*** (0.009) in column (1); 0.339*** (0.012) column (2); 0.340*** (0.025) column (3); 0.352*** (0.013) column (4)).
    - Consumption elasticity with respect to home value (housing wealth): ranges from 0.1 to 0.15 and is higher in post-COVID (e.g., 0.100*** (0.006) column (1); 0.153*** (0.018) column (4); 0.180*** (0.013) column (7)).
    - Savings and stocks elasticities are modest and not consistently significant (e.g., Savings 0.003*** (0.000) column (1); Stocks 0.003*** (0.000) column (1)).
    - Cumulative saving since Covid (lagged) shows growth in importance: L.Cumulative saving since Covid coefficients: 0.004 (0.003) in pre-COVID proxy; 0.024** (0.009) COVID; 0.052*** (0.011) post-COVID in specific specifications.
  - When income is split into non-asset income, current transfers, and income from assets:
    - Non-asset income elasticities slightly lower than total income.
    - Current transfers generally positively correlated with consumption except during COVID period.
    - Dividend/interest/rent elasticities positive and significant (e.g., 0.018*** (0.001) column (1); 0.013*** (0.003) column (6); 0.012*** (0.002) column (7)).
- Heterogeneity by home-ownership status (Equation 3; Table 3 summary):
  - Household groups: homeowners with mortgage, homeowners without mortgage, renters, occupants without rental payment.
  - Income elasticities:
    - Higher for homeowners and renters, around 0.4, with marginal increase post-COVID.
    - Occupants without rental payments: income elasticities decline significantly and are lowest and statistically insignificant.
  - Housing wealth elasticities:
    - Higher for homeowners with mortgage at about 0.2.
    - Post-COVID increases more pronounced for homeowners without mortgage, nearly doubling from 0.07 to 0.13.
  - Cumulative savings elasticities:
    - Rise post-COVID especially for homeowners, followed by renters.
- Interpretation:
  - Better-off households (less constrained) experienced accumulation of more income and housing wealth and saw increases in elasticities post-COVID, implying a potentially large contribution of these groups to overall consumption growth.

*Source: wpiea2024128-print-pdf - Section 2*

### Section 3

### Section 3 — Household-level and Aggregate Drivers of Consumption

### Key household-level elasticities and group heterogeneity
- Consumption-income elasticities (selected):
  - Homeowner w/ mortgage: 0.392*** (Pre-COVID), 0.450*** (COVID), 0.393*** (Post-COVID) — reported in Table 3 (Consumption expenditure; log).
  - Homeowner w/o mortgage: 0.424*** (Pre-COVID), 0.382*** (COVID), 0.441*** (Post-COVID).
  - Renter: 0.429*** (Pre-COVID), 0.443*** (COVID), 0.443*** (Post-COVID).
  - Occupant w/o rent payment: 0.363*** (Pre-COVID), 0.574*** (COVID), 0.069 (Post-COVID).
- Income elasticities excluding transfers and dividend/interest/rent (selected):
  - Homeowner w/ mortgage: 0.325*** (Pre-COVID), 0.409*** (COVID), 0.358*** (Post-COVID).
  - Homeowner w/o mortgage: 0.286*** (Pre-COVID), 0.335*** (COVID), 0.317*** (Post-COVID).
  - Renter: 0.348*** (Pre-COVID), 0.340*** (COVID), 0.367*** (Post-COVID).
- Housing value elasticities (selected):
  - Homeowner w/ mortgage: 0.166*** (Pre-COVID), 0.176*** (COVID), 0.214*** (Post-COVID).
  - Homeowner w/o mortgage: 0.072*** (Pre-COVID), 0.114*** (COVID), 0.149*** (Post-COVID).
- Elasticities with respect to cumulative implied saving since COVID (selected):
  - Homeowner w/ mortgage: 0.011*** (Pre-COVID), 0.035*** (COVID), 0.056*** (Post-COVID).
  - Homeowner w/o mortgage: 0.006 (Pre-COVID), 0.039*** (COVID), 0.107*** (Post-COVID).
  - Renter: -0.006 (Pre-COVID), 0.042*** (COVID), 0.039*** (Post-COVID).

Notes: cohort and time fixed effects included; samples are Pre-COVID: 2013Q1-2019Q4; COVID: 2020Q1-2021Q1; Post-COVID: 2021Q2-2022Q4. Robust standard errors in parentheses. Significance: * p<0.1; ** p<0.05; *** p<0.01.

### Saving marginal propensities to consume (MPCs) by income
- Method: MPCs derived by multiplying estimated elasticities with the consumption-to-savings ratio of each household income quintile.
- Main finding:
  - Saving MPCs are highest among lower-income households, indicating faster depletion of excess savings since COVID.
  - Average behavior: low-income households spend approximately 40-45 cents for every dollar increase in cumulative savings.
  - Top 40 percent of households by income: response of about 20-25 cents for every dollar increase in cumulative savings.

### Heterogeneous consumption responses across expenditure categories
- Consumption divided into five non-overlapping groups: recreational expenditure; automobile purchase; healthcare; home food; others.
- Broad patterns:
  - Higher income elasticities for recreation, automobile, and healthcare; lower elasticity for home food.
  - Homeowners and renters generally show higher income elasticities across categories, except home food where occupants without rent payment exhibit the highest elasticities.
  - Post-COVID increases in income elasticities are particularly pronounced for recreation expenditures.
- Selected category-specific elasticities from Table 4 (Pre-COVID and Post-COVID shown where available):
  - Recreation (Income after tax excl. transfers and dividend/interest/rent; log #):
    - Homeowner w/ mortgage: 0.534*** (Pre-COVID), 0.655*** (Post-COVID).
    - Homeowner w/o mortgage: 0.453*** (Pre-COVID), 0.653*** (Post-COVID).
    - Renter: 0.561*** (Pre-COVID), 0.762*** (Post-COVID).
    - Occupant w/o rent payment: 0.221*** (Pre-COVID), 0.423*** (Post-COVID).
  - Home value elasticities for Recreation:
    - Homeowner w/ mortgage: 0.208*** (Pre-COVID), 0.214*** (Post-COVID).
    - Homeowner w/o mortgage: 0.177*** (Pre-COVID), 0.328*** (Post-COVID).
  - L.Cumulative implied saving since COVID (Recreation):
    - Homeowner w/ mortgage: 0.019 (Pre-COVID), 0.151*** (Post-COVID).
    - Homeowner w/o mortgage: -0.002 (Pre-COVID), 0.175*** (Post-COVID).
    - Renter: -0.019 (Pre-COVID), 0.062 (Post-COVID).
  - Illustrative post-COVID magnitudes highlighted in text:
    - Post-COVID elasticities for recreational spending among homeowners without mortgage are estimated at 0.3 with respect to housing wealth and approximately 0.2 with respect to excess saving (text discussion summarizing Table 4 results).

### Robustness checks
- Including other financial assets (gross and net of liabilities):
  - Does not change housing wealth elasticities; elasticities with respect to other financial assets are notably small.
  - Two interpretations noted: (1) housing wealth elasticities represent genuine effect; or (2) they reflect a blend of housing and non-housing wealth effects (potential measurement/reporting errors).
  - Correlations between housing and financial assets are weaker in this data (CE) than in the Survey of Consumer Finance, likely due to under-reported financial assets in the CE.
- Controlling for fixed-rate mortgage and student loans:
  - Households with fixed-rate mortgage consume slightly more; coefficients become insignificant during post-COVID period and in specifications controlling for cumulative saving.
  - No statistically significant relationship between consumption and student loan status; interpret with caution due to possible underreporting and nonlinear debt-amount effects.

### Aggregate decomposition: how much of consumption growth is explained by micro drivers
- Method: apply household-level elasticities (last three columns of Table 2) to macro-level changes in after-tax wages and salary income, current transfers, dividend/interest and rental income, and excess pandemic saving drawdowns. Macro data from BEA and Federal Reserve Board; deflated with the PCE price index.
- Key aggregate magnitudes and contributions (expressed as percent deviation from 2019Q4 level):
  - Trough in 2020Q2: consumption was 10 percent below the 2019Q4 level; loss in real incomes explained only a small share; residual dominated due to mobility restrictions and supply shocks.
  - By end-2022: aggregate consumption around 7 percent above pre-pandemic level, decomposed as:
    - 4.6 percent explained by the rise in housing wealth (post-COVID household-level estimates).
    - 1.6 percent explained by higher real household incomes.
    - 1.1 percent explained by the drawdown of excess savings from the pandemic.
  - By end-2023 (additional increase relative to 2022):
    - Higher real incomes contributed an additional 1.1 percent (computed as 2.7 − 1.6 = 1.1percent).
    - Housing wealth contributed an additional 0.6 percent (computed as 5.2 − 4.6 = 0.6 percent).
    - Excess saving drawdown contributed 0.5 percent relative to end-2019 (noting drawdown slowed as households exhausted pandemic savings).
  - One fifth of the consumption growth in 2023Q4 relative to 2019Q4 remains unexplained (grey residual).
- Cumulative draw-down of excess saving implied by estimates:
  - Sum of yellow bars equivalent to about 2 Trn Dollar (in nominal terms), noted to be very close to Abdelrahman and Oliveira (2023a)’s estimate of 2.1 Trn Dollar over roughly the same period.
- External validity checks:
  - Residuals largest and negative during lockdown; close to zero when consumption completed recovery; positive but small in 2023Q4 when consumption unusually strong.
  - Micro estimates applied out-of-sample (2023) explain around 80 percent of observed aggregate consumption growth in 2023.

### Main takeaways and outlook
- Primary finding: housing wealth has become the most important driver of private consumption in recent years; this is consistent across state-level and household-level analyses.
- Distributional pattern: consumption growth concentrated in households in upper part of income distribution.
- Reasons for larger post-pandemic propensity to consume out of housing wealth (hypotheses for future research):
  - Housing wealth is more widely held across segments (including larger shares of home ownership among the young), potentially raising aggregate consumption.
  - Estimated housing wealth effect may partially capture under-reported non-housing financial wealth that is correlated with housing wealth (Survey of Consumer Finances shows homeowners hold much larger non-housing financial wealth and that it has appreciated strongly).
  - Related evidence: cryptocurrency wealth correlated with housing wealth and associated with much larger MPCs in other work (Aiello et al. (2023) referenced).
- Policy-relevant implications:
  - Continued disinflation, resilient labor market, and robust wage growth would support consumption via higher real incomes.
  - Continued housing price appreciation (as interest rates decline and demand rises) could provide further tailwind to U.S. consumption and growth because households are now more responsive to wealth changes.
  - Conversely, any repricing of housing and other financial assets poses downside risks to household consumption and growth — a sharp wealth retrenchment could lead to notable consumption weakness.

*Source: wpiea2024128-print-pdf — Section 3 (Household-level results, spending categories, robustness checks, and aggregation exercise).*

### Section 4

### wpiea2024128-print-pdf - Section 4

### References cited
- Aiello, Darren, Scott R Baker, Tetyana Balyuk, Marco Di Maggio, Mark J Johnson, and Jason D Kotter.2023. “The effects of cryptocurrency wealth on household consumption and investment.” National Bureau of Economic Research.
- Beraja, Martin, and Christian K Wolf.2021. “Demand composition and the strength of recoveries.” National Bureau of Economic Research.
- Broda, Christian, and Jonathan A Parker.2014. “The economic stimulus payments of 2008 and the aggregate demand for consumption.”Journal of Monetary Economics, 68: S20–S36.
- Browning, Martin, Angus Deaton, and Margaret Irish.1985. “A Profitable Approach to Labor Supply and Commodity Demands over the Life-Cycle.”Econometrica, 53(3): 503–43.
- Caceres, Carlos.2019.Analyzing the effects of financial and housing wealth on consumption using micro data.International Monetary Fund.
- Carroll, Christopher D.2001. “A theory of the consumption function, with and without liquidity constraints.”Journal of Economic perspectives, 15(3): 23–45.
- Carroll, Christopher, Jiri Slacalek, Kiichi Tokuoka, and Matthew N. White.2017. “The distribution of wealth and the marginal propensity to consume.”Quantitative Economics, 8(3): 977–1020.
- Casado, Miguel Garza, Britta Glennon, Julia Lane, David McQuown, Daniel Rich, and Bruce A. Weinberg.2020. “The Aggregate Effects of Fiscal Stimulus: Evidence from the COVID-19 Unemployment Supplement.” National Bureau of Economic Research, Inc NBER Working Papers 27576.
- Case, Karl E., John M. Quigley, and Robert J. Shiller.2005. “Comparing Wealth Effects: The Stock Market versus the Housing Market.”The B.E. Journal of Macroeconomics, 5(1): 1–34.
- Cho, Yunho, James Morley, and Aarti Singh.2024. “Did marginal propensities to consume change with the housing boom and bust?”Journal of Applied Econometrics, 39(1): 174–199.
- Deaton, Angus.1985. “Panel data from time series of cross-sections.”Journal of Econometrics, 30(1): 109–126.
- Fisher, Jonathan D., David S. Johnson, Timothy M. Smeeding, and Jeffrey P. Thompson.2020. “Estimating the marginal propensity to consume using the distributions of income, consumption, and wealth.”Journal of Macroeconomics, 65(C).
- Gamber, William, James Graham, and Anirudh Yadav.2023. “Stuck at home: Housing demand during the COVID-19 pandemic.”Journal of Housing Economics, 59: 101908.
- Guren, Adam M, Alisdair McKay, Emi Nakamura, and Jón Steinsson.2021. “Housing wealth effects: The long view.”The Review of Economic Studies, 88(2): 669–707.
- Jappelli, Tullio, and Luigi Pistaferri.2010. “The Consumption Response to Income Changes.” National Bureau of Economic Research, Inc NBER Working Papers 15739.
- Johnson, David S., Jonathan A. Parker, and Nicholas S. Souleles.2006. “Household Expenditure and the Income Tax Rebates of 2001.”American Economic Review, 96(5): 1589–1610.
- Kaplan, Greg, Giovanni L. Violante, and Justin Weidner.2014. “The Wealthy Hand-to-Mouth.”Brookings Papers on Economic Activity, 45(1 (Spring): 77–153.
- Kent, Ana H., and Lowell R. Ricketts.2024. “Millennials and Older Gen Z’ers Made Significant Wealth Gains in 2022.” Federal Reserve Bank of St. Louis.
- Mian, Atif, Kamalesh Rao, and Amir Sufi.2013. “Household balance sheets, consumption, and the economic slump.”The Quarterly Journal of Economics, 128(4): 1687–1726.
- Parker, Jonathan A., Jake Schild, Laura Erhard, and David Johnson.2022. “Economic Impact Payments and Household Spending During the Pandemic.” National Bureau of Economic Research, Inc NBER Working Papers 30596.
- R. Baker, Scott, Robert A Farrokhnia, Steffen Meyer, Michaela Pagel, and Constantine Yannelis.2023. “Income, Liquidity, and the Consumption Response to the 2020 Economic Stimulus Payments.”Review of Finance, 27(6): 2271–2304.

### Appendix — Figure A1: Real Consumption and Retail Employment, pre- vs. post-COVID
- Figure compares ln real consumption (demeaned, detrended) and ln retail employment (demeaned, detrended).
- Two samples plotted:
  - 2015-2019 (pre-COVID): blue squares.
  - 2020-2022 (post-COVID): red dots.
- Both variables are in logarithm, residualized by regressing on state and year fixed effects.
- The within R2 for both regressions are around 0.95.
- Axis markers/values shown: -.1, -.05, 0, .05, .1 for ln real consumption; -.15, -.1, -.05, 0, .05, .1 for ln retail employment.
- Figure includes a 45-degree reference line.

### Appendix — Table A1: Panel unit root test (H0: panels contain unit roots; Ha: panels are stationary)
- Reported p-value
  - Retail employment: 0.0 (Pre-COVID) 0.0 (Post-COVID)
  - Income: 0.1 (Pre-COVID) 0.0 (Post-COVID)
  - Wage: 0.0 (Pre-COVID) 0.0 (Post-COVID)
  - Dividend/interest/rent: 0.0 (Pre-COVID) 0.0 (Post-COVID)
  - Non-wage income: 0.1 (Pre-COVID) 0.0 (Post-COVID)
  - Housing wealth: 0.0 (Pre-COVID) 0.0 (Post-COVID)
- Note: Levin–Lin–Chu panel unit root test of the state-level variables (in logarithm and demeaned). P-value is reported.

### Appendix — Table A2: State-level Results Using Imputed Real Consumption
- Dependent variable: real consumption (log, imputed by historical relationship between real consumption and retail employment).
- Samples:
  - Column (1): 1990Q1-2019Q4 (Observations: 6,120; Within R-squared: 0.663; Quarter FE: Y; State FE: Y)
  - Columns (2)-(4): 20Q1-23Q4 (Observations: 816; Within R-squared: 0.176, 0.189, 0.270 respectively; Quarter FE: Y; State FE: Y)
- Coefficients and standard errors (clustered at state level):
  - Wage: 0.1762*** (0.048) [Column (1)]
  - Asset-based income (dividend/interest/rent):
    - 0.1089** (0.044) [Col (1)]
    - 0.0602* (0.033) [Col (2)]
    - 0.0238 (0.037) [Col (3)]
    - 0.0318 (0.036) [Col (4)]
  - Income other than dividend/interest/rent:
    - 0.3286*** (0.072) [Col (1)]
    - 0.0031 (0.034) [Col (2)]
    - -0.0210 (0.037) [Col (3)]
  - Income other than wage and dividend/interest/rent:
    - -0.0386** (0.015) [Col (4)]
  - Housing wealth:
    - 0.0379 (0.026) [Col (1)]
    - 0.0944*** (0.012) [Col (2)]
    - 0.0872*** (0.012) [Col (3)]
    - 0.0572*** (0.012) [Col (4)]
  - Cumulative income since COVID, 1-quarter lag:
    - 0.1891** (0.083) [Col (3)]
    - 0.0644 (0.090) [Col (4)]
  - Mobility, retail/transit/work:
    - 0.0004*** (0.000) [Col (2)]
    - 0.0005*** (0.000) [Col (3)]
    - 0.0004*** (0.000) [Col (4)]
- Significance notation: *, ** and *** represent significance level at 10, 5 and 1%, respectively.
- Note: Column descriptions — (1) pre-COVID sample; (2) post-COVID sample; (3) adds lagged cumulative income since COVID; (4) separates wage out from other income.

### Appendix — Table A3: Household-level Results: Robustness Checks - Other Financial Assets
- Dependent variable: Consumption expenditure; log.
- Samples by column grouping: Pre-COVID, COVID, Post-COVID (columns (1)-(3)) and repeated (4)-(6).
- Observations:
  - Pre-COVID: 17094
  - COVID: 2423
  - Post-COVID: 4995
  - (Repeated totals appear for columns (4)-(6): 17094, 2423, 4995)
- R2:
  - 0.805, 0.872, 0.839 (for first three columns)
  - 0.805, 0.871, 0.839 (for next three columns)
- Coefficients (standard errors in parentheses):
  - Income after tax excl. transfers/dividend/interest/rent; log:
    - 0.299*** (0.016)
    - 0.348*** (0.037)
    - 0.335*** (0.021)
    - 0.299*** (0.016)
    - 0.350*** (0.037)
    - 0.335*** (0.021)
  - Home value; log:
    - 0.094*** (0.009)
    - 0.136*** (0.024)
    - 0.181*** (0.013)
    - 0.094*** (0.009)
    - 0.136*** (0.024)
    - 0.180*** (0.013)
  - Savings; log:
    - 0.002** (0.001)
    - -0.002 (0.002)
    - 0.002* (0.001)
    - 0.002*** (0.001)
    - -0.001 (0.002)
    - 0.002* (0.001)
  - 1{Zero current transfers}:
    - 0.030*** (0.009)
    - -0.005 (0.029)
    - 0.026 (0.018)
    - 0.030*** (0.009)
    - -0.008 (0.029)
    - 0.026 (0.018)
  - Current transfer; log:
    - 0.006*** (0.002)
    - -0.000 (0.005)
    - 0.005* (0.003)
    - 0.006*** (0.002)
    - -0.001 (0.005)
    - 0.005* (0.003)
  - Dividend/interest/rent; log:
    - 0.018*** (0.001)
    - 0.012*** (0.003)
    - 0.012*** (0.002)
    - 0.018*** (0.001)
    - 0.013*** (0.003)
    - 0.012*** (0.002)
  - L.Cumulative saving since COVID; log:
    - 0.004 (0.003)
    - 0.024** (0.009)
    - 0.052*** (0.011)
    - 0.004 (0.003)
    - 0.024** (0.009)
    - 0.052*** (0.011)
  - Financial assets (excl. savings); log:
    - 0.002*** (0.001)
    - 0.004*** (0.002)
    - 0.001 (0.001)
  - Net financial assets (excl. savings); log:
    - 0.002*** (0.001)
    - 0.003** (0.001)
    - 0.001 (0.001)
- Controls: Cohort and time fixed effects; household characteristics included.
- Sample definitions:
  - Pre-COVID sample: 2013Q1-2019Q4
  - COVID sample: 2020Q1-2021Q1
  - Post-COVID sample: 2021Q2-2022Q4
- Note: 1/ * p<0.1; ** p<0.05; and *** p<0.01. 4/ Robust standard errors in parentheses.

### Appendix — Table A4: Household-level Results: Robustness Checks - Fixed Rate Mortgage and Student Loans
- Dependent variable: Consumption expenditure; log.
- Samples across columns (1)-(9): Pre-COVID, COVID, Post-COVID repeated thrice.
- Observations per sample grouping:
  - Pre-COVID: 17094
  - COVID: 2423
  - Post-COVID: 4995
  - (Repeated across column groupings yields same totals: 17094, 2423, 4995)
- R2:
  - 0.805, 0.871, 0.839 repeated for groups of three columns.
- Selected coefficients (standard errors in parentheses):
  - Income after tax excl. transfers/dividend/interest/rent; log:
    - 0.300*** (0.016)
    - 0.352*** (0.037)
    - 0.335*** (0.021)
    - 0.299*** (0.016)
    - 0.349*** (0.037)
    - 0.335*** (0.021)
    - 0.299*** (0.016)
    - 0.350*** (0.037)
    - 0.335*** (0.021)
  - Home value; log:
    - 0.094*** (0.009)
    - 0.135*** (0.024)
    - 0.180*** (0.013)
    - 0.094*** (0.009)
    - 0.136*** (0.024)
    - 0.181*** (0.014)
    - 0.094*** (0.009)
    - 0.136*** (0.024)
    - 0.181*** (0.013)
  - Savings; log:
    - 0.003*** (0.001)
    - 0.000 (0.002)
    - 0.002** (0.001)
    - 0.002** (0.001)
    - -0.002 (0.002)
    - 0.002 (0.001)
    - 0.002** (0.001)
    - -0.001 (0.002)
    - 0.002 (0.001)
  - Stocks; log:
    - 0.002*** (0.001)
    - 0.002 (0.002)
    - 0.002 (0.001)
  - 1{Zero current transfers}:
    - 0.030*** (0.009)
    - -0.007 (0.029)
    - 0.026 (0.018)
    - 0.030*** (0.009)
    - -0.004 (0.029)
    - 0.026 (0.018)
    - 0.030*** (0.009)
    - -0.007 (0.029)
    - 0.026 (0.018)
  - Current transfer; log:
    - 0.006*** (0.002)
    - -0.000 (0.005)
    - 0.005* (0.003)
    - 0.005*** (0.002)
    - -0.000 (0.005)
    - 0.005* (0.003)
    - 0.006*** (0.002)
    - -0.000 (0.005)
    - 0.005* (0.003)
  - Dividend/interest/rent; log:
    - 0.018*** (0.001)
    - 0.013*** (0.003)
    - 0.012*** (0.002)
    - 0.018*** (0.001)
    - 0.012*** (0.003)
    - 0.012*** (0.002)
    - 0.018*** (0.001)
    - 0.013*** (0.003)
    - 0.012*** (0.002)
  - L.Cumulative saving since COVID; log:
    - 0.004 (0.003)
    - 0.024*** (0.009)
    - 0.052*** (0.011)
    - 0.004 (0.003)
    - 0.024*** (0.009)
    - 0.052*** (0.011)
    - 0.004 (0.003)
    - 0.025*** (0.009)
    - 0.052*** (0.011)
  - Financial assets (excl. savings); log:
    - 0.002*** (0.001)
    - 0.004*** (0.002)
    - 0.001 (0.001)
  - Net financial assets (excl. savings); log:
    - 0.002*** (0.001)
    - 0.003** (0.001)
    - 0.001 (0.001)
  - 1{Fixed rate mortgage}:
    - 0.028 (0.024)
    - 0.124* (0.075)
    - -0.049 (0.050)
    - 0.027 (0.024)
    - 0.129* (0.076)
    - -0.049 (0.050)
    - 0.027 (0.024)
    - 0.129* (0.076)
    - -0.049 (0.050)
  - 1{Student loans}:
    - 0.038 (0.026)
    - 0.064 (0.055)
    - -0.006 (0.042)
    - 0.031 (0.026)
    - 0.057 (0.054)
    - -0.007 (0.042)
    - 0.046* (0.026)
    - 0.081 (0.055)
    - -0.001 (0.042)
- Controls: Cohort and time fixed effects; household characteristics included.
- Sample definitions:
  - Pre-COVID sample: 2013Q1-2019Q4
  - COVID sample: 2020Q1-2021Q1
  - Post-COVID sample: 2021Q2-2022Q4
- Note: 1/ * p<0.1; ** p<0.05; and *** p<0.01. 4/ Robust standard errors in parentheses.

*Content derived from wpiea2024128-print-pdf - Section 4*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024128-print-pdf.pdf_