## wpiea2019115

## Source details

**Canonical URL:** [wpiea2019115](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019115.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019115.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019115.pdf.json)

---

### I. Introduction: motivation and scope
- Financial asset and house prices have experienced large swings; real U.S. house prices fell by over a third between 2006 and 2012.
- Consumption accounts for almost 70 percent of total output in the U.S., and an average 1.6 percent contribution to annual GDP growth over the last decade.
- Central question: do households base consumption partly on the value of their net assets (wealth)? Specifically, what is the marginal propensity to consume (MPC) out of net wealth, and do MPCs differ by wealth type and household characteristics?
- This paper uses micro (household-level) longitudinal data to estimate MPCs and to identify determinants of heterogeneity.

### II. Data and empirical methodology
- Data:
  - Panel Study of Income Dynamics (PSID) longitudinal survey.
  - Sample period: 1999 to 2017, every other year.
  - Sample size: approximately 7,000-9,000 households depending on year; total observations reported in regressions: 83,589 (and similar counts across tables).
  - All nominal values converted to 2017 real dollars using the personal consumption expenditure (PCE) price index.
- Key measures:
  - Wealth components: primary home (net housing wealth = home value minus mortgage), other real estate, farm/business, motor vehicles, bank accounts and similar, stock holdings (excluding IRAs), other savings, IRAs; total net wealth = sum of eight wealth types minus debt.
  - Consumption = detailed PSID household expenditure items (food, housing-related, transportation, education, childcare, healthcare, clothing, vacation and recreation).
- Micro estimation:
  - Core regression: C_{i,t} = α + ρ I_{i,t} + β W_{i,t} + γ Z_{i,t} + ε_{i,t}, where β is the MPC out of wealth W.
  - β also modeled as function of household characteristics H_{i,t} via interactions: β_{i,t} = Φ(H_{i,t}).
  - Core estimations omit household fixed effects to preserve variation for estimating interactions with persistent characteristics.
- Macro estimation:
  - VAR on quarterly real private consumption, S&P-500 stock price index, Case-Shiller house price index, and aggregate household disposable income.
  - Impulse response functions used to assess consumption sensitivity to 10 percent stock and house price shocks.

### III. Macro-based evidence (aggregate VAR results)
- Impulse responses to a 10 percent shock:
  - House price shock: larger and longer-lasting effect on real private consumption than a commensurate stock price shock; consumption response increases between the first and third year after the shock and remains significant over multiple years (confidence bands).
  - Stock price shock: consumption response statistically significant four quarters after the shock but not significant after 3 years (based on 5 percent confidence bands).
- Implication: changes in house prices are more important for aggregate consumption than stock price movements.

### IV. Micro-based evidence: main results

- Core estimated MPCs (controlling for household income; estimation period: 1999-2017; observations typically 83,589):
  - Net home equity (net housing wealth): 0.043*** (robust standard error (0.003)). Interpretation in text: about 4 cents increase in consumption per additional dollar of housing equity.
  - Gross home assets: 0.062*** (0.003).
  - Gross home-related debt: 0.131*** (0.004).
  - Net nonfinancial assets: 0.003*** (0.001).
  - Gross nonfinancial assets: 0.004*** (0.001).
  - Financial assets: 0.003** (0.001).
  - Stock holdings: 0.002 (0.002) — not statistically significant in core model.
  - Total net wealth: 0.003*** (0.001).
  - Total gross wealth: 0.004*** (0.001).
  - Net wealth, excluding home equity: 0.002*** (0.001).
  - Gross wealth, excluding home equity: 0.002*** (0.001).

- Role of housing channel:
  - Gross housing wealth MPC exceeds net housing MPC, and the large coefficient on gross home-related debt (0.131***) suggests a borrowing/liquidity channel where households increase home-related debt to support consumption following housing wealth gains.

- MPCs over time and across states:
  - Time variation:
    - Housing MPCs increased around the turn of the century and through the early 2000s, but have trended broadly downward over the past 12 years (1999–2017), with some volatility.
    - Stock-holdings MPCs are close to zero and generally not statistically significant across most periods.
  - State variation:
    - Significant housing MPCs observed in several states (e.g., Georgia, Massachusetts, New York, Vermont, Arkansas, New Mexico, Texas).
    - Stock-holdings MPCs positive and significant in only a few states (Northeast and some southern states); magnitudes exceeding 0.05 (5 cents) observed only in Vermont and Wyoming for stock holdings.
    - Overall: housing wealth more important than other forms of wealth for consumption across states.

- Heterogeneity in MPCs by household characteristics:
  - Homeownership:
    - Housing wealth effects operate only through homeowners; no effect for renters.
  - Age:
    - Younger households exhibit larger housing MPCs. Example: ages 25-44 have almost 0.08 MPC (about 8 cents) per dollar of net housing wealth in some specifications versus less than 0.04 for those aged > 65.
  - Employment status:
    - Working households have larger MPCs than non-working households (retired, unemployed, students, disabled).
  - Education and marital status:
    - Those who did not attend college tend to have higher MPCs than college-educated; married individuals tend to have higher MPCs than those not married.
  - Life satisfaction:
    - Households reporting higher life satisfaction have higher MPCs than those reporting lower life satisfaction.
  - Income distribution and MPCs:
    - Simple interaction: housing MPC falls as household income rises.
    - Full analysis: non-monotonic relationship — MPCs for net housing wealth are relatively larger at both ends of the income distribution (lowest and highest income groups) compared to middle-income groups.
    - Lowest-income households: higher MPC possibly because they spend out of non-income sources and may borrow or dispose assets; sample caveat: lowest decile includes households reporting zero or negative income.
    - Highest-income households (above the 90th percentile): higher MPCs out of gross housing assets and higher home-related debt MPCs, possibly reflecting greater access to credit allowing equity extraction.
  - Liquidity proxy:
    - Households with higher cash-to-income ratios (more liquid) have lower housing MPCs; borrowers/liquidity-constrained households exhibit higher MPCs.
  - Total net wealth:
    - MPCs for total net wealth generally lower for higher-income households; magnitudes are small (e.g., 0.003*** in core), consistent with stock MPCs being small since stocks are concentrated among higher-income households.

- Implications of demographic changes:
  - Demographic trends (PSID and U.S. Census projections) point to population aging, declining homeownership and changing educational and marital status shares.
  - Estimated direct effect: population aging is estimated to lower the housing MPC by about 0.3 cents out of every additional dollar in net home wealth every decade.
  - Overall implication: underlying demographic trends point to potentially lower aggregate MPCs out of wealth going forward.

### V. Robustness checks (appendix summary)
- Robustness across specifications:
  - Results for net housing wealth MPC (0.043***) robust to addition of household-specific controls, time fixed effects, household fixed effects, and composition adjustments (Table A1).
  - Stock holdings MPCs remain small and not significant across robustness checks (Table A2).
  - Total net wealth MPCs (0.003***) robust across alternative specifications (Table A3).
- Additional robustness:
  - Deflating using CPI rather than PCE leaves estimation results virtually unchanged.
  - Truncated samples excluding lowest and highest 1 percent of income distribution examined; interaction terms and group dummies used to probe non-linearities.

### VI. Quantitative implications and conclusion
- Representative quantitative results:
  - Net housing wealth MPC: 0.043***.
  - Gross home assets MPC: 0.062***.
  - Gross home-related debt MPC: 0.131***.
  - Total net wealth MPC: 0.003***.
  - Stock holdings MPC: 0.002 (not statistically significant in core estimates).
- Aggregate implication:
  - Using estimated MPCs across income groups, a sustained 10 percent fall in house prices would lead to a fall in aggregate consumption of about 1-1.4 percent.
  - Recent increases in financial asset values appear to have done little to support aggregate consumption because MPCs out of financial and stock wealth are small or insignificant.
- Policy-relevant inferences:
  - Housing wealth fluctuations have meaningful consumption effects, amplified by borrowing/collateral channels.
  - Heterogeneity by age, income, employment, and credit access matters for transmission of house-price shocks to consumption.
  - Demographic trends (aging) are likely to dampen the sensitivity of consumption to wealth going forward.

*Source: IMF staff calculations based on the Panel Study of Income Dynamics (PSID) and aggregate time series; estimation period 1999–2017 as reported in wpiea2019115.*

### References .............................................................................................................

### wpiea2019115 - References

### I. Introduction: motivation and scope
- Financial asset and house prices have experienced large swings; real U.S. house prices fell by over a third between 2006 and 2012.
- Consumption accounts for almost 70 percent of total output in the U.S., and an average 1.6 percent contribution to annual GDP growth over the last decade.
- Central question: do households base consumption partly on the value of their net assets (wealth)? Specifically, what is the marginal propensity to consume (MPC) out of net wealth, and do MPCs differ by wealth type and household characteristics?
- This paper uses micro (household-level) longitudinal data to estimate MPCs and to identify determinants of heterogeneity.

### II. Data and empirical methodology
- Data:
  - Panel Study of Income Dynamics (PSID) longitudinal survey.
  - Sample period: 1999 to 2017, every other year.
  - Sample size: approximately 7,000-9,000 households depending on year; total observations reported in regressions: 83,589 (and similar counts across tables).
  - All nominal values converted to 2017 real dollars using the personal consumption expenditure (PCE) price index.
- Key measures:
  - Wealth components: primary home (net housing wealth = home value minus mortgage), other real estate, farm/business, motor vehicles, bank accounts and similar, stock holdings (excluding IRAs), other savings, IRAs; total net wealth = sum of eight wealth types minus debt.
  - Consumption = detailed PSID household expenditure items (food, housing-related, transportation, education, childcare, healthcare, clothing, vacation and recreation).
- Micro estimation:
  - Core regression: C_{i,t} = α + ρ I_{i,t} + β W_{i,t} + γ Z_{i,t} + ε_{i,t}, where β is the MPC out of wealth W.
  - β also modeled as function of household characteristics H_{i,t} via interactions: β_{i,t} = Φ(H_{i,t}).
  - Core estimations omit household fixed effects to preserve variation for estimating interactions with persistent characteristics.
- Macro estimation:
  - VAR on quarterly real private consumption, S&P-500 stock price index, Case-Shiller house price index, and aggregate household disposable income.
  - Impulse response functions used to assess consumption sensitivity to 10 percent stock and house price shocks.

### III. Macro-based evidence (aggregate VAR results)
- Impulse responses to a 10 percent shock:
  - House price shock: larger and longer-lasting effect on real private consumption than a commensurate stock price shock; consumption response increases between the first and third year after the shock and remains significant over multiple years (confidence bands).
  - Stock price shock: consumption response statistically significant four quarters after the shock but not significant after 3 years (based on 5 percent confidence bands).
- Implication: changes in house prices are more important for aggregate consumption than stock price movements.

### IV. Micro-based evidence: main results

H3: Estimated MPCs (household-level regressions)
- Core estimated MPCs (controlling for household income; estimation period: 1999-2017; observations typically 83,589):
  - Net home equity (net housing wealth): 0.043*** (robust standard error (0.003)). Interpretation in text: about 4 cents increase in consumption per additional dollar of housing equity.
  - Gross home assets: 0.062*** (0.003).
  - Gross home-related debt: 0.131*** (0.004).
  - Net nonfinancial assets: 0.003*** (0.001).
  - Gross nonfinancial assets: 0.004*** (0.001).
  - Financial assets: 0.003** (0.001).
  - Stock holdings: 0.002 (0.002) — not statistically significant in core model.
  - Total net wealth: 0.003*** (0.001).
  - Total gross wealth: 0.004*** (0.001).
  - Net wealth, excluding home equity: 0.002*** (0.001).
  - Gross wealth, excluding home equity: 0.002*** (0.001).
- Role of housing channel:
  - Gross housing wealth MPC exceeds net housing MPC, and the large coefficient on gross home-related debt (0.131***) suggests a borrowing/liquidity channel where households increase home-related debt to support consumption following housing wealth gains.

H3: MPCs over time and across states
- Time variation:
  - Housing MPCs increased around the turn of the century and through the early 2000s, but have trended broadly downward over the past 12 years (1999–2017), with some volatility.
  - Stock-holdings MPCs are close to zero and generally not statistically significant across most periods.
- State variation:
  - Significant housing MPCs observed in several states (e.g., Georgia, Massachusetts, New York, Vermont, Arkansas, New Mexico, Texas).
  - Stock-holdings MPCs positive and significant in only a few states (Northeast and some southern states); magnitudes exceeding 0.05 (5 cents) observed only in Vermont and Wyoming for stock holdings.
  - Overall: housing wealth more important than other forms of wealth for consumption across states.

H3: The role of household characteristics (heterogeneity in MPCs)
- Method: interact wealth measures with household-group dummies (homeownership, age groups, employment status, education, marital status, income groups).
- Key findings on net housing wealth MPC heterogeneity:
  - Homeowners vs renters:
    - Housing wealth effects operate only through homeowners; no effect for renters.
  - Age:
    - Younger households exhibit larger housing MPCs. Example: ages 25-44 have almost 0.08 MPC (about 8 cents) per dollar of net housing wealth in some specifications versus less than 0.04 for those aged > 65.
  - Employment status:
    - Working households have larger MPCs than non-working households (retired, unemployed, students, disabled).
  - Education and marital status:
    - Those who did not attend college tend to have higher MPCs than college-educated; married individuals tend to have higher MPCs than those not married.
  - Life satisfaction:
    - Households reporting higher life satisfaction have higher MPCs than those reporting lower life satisfaction.
- Income distribution and MPCs:
  - Interaction regressions and income-group interactions indicate housing MPC falls as household income rises (simple interaction), but full analysis shows a non-monotonic relationship:
    - MPCs for net housing wealth are relatively larger at both ends of the income distribution (lowest and highest income groups) compared to middle-income groups.
    - Lowest-income households: higher MPC possibly because they spend out of non-income sources and may borrow or dispose assets; sample caveat: lowest decile includes households reporting zero or negative income.
    - Highest-income households (above the 90th percentile): higher MPCs out of gross housing assets and higher home-related debt MPCs, possibly reflecting greater access to credit allowing equity extraction.
  - Liquidity proxy:
    - Households with higher cash-to-income ratios (more liquid) have lower housing MPCs; borrowers/liquidity-constrained households exhibit higher MPCs.
- Total net wealth:
  - MPCs for total net wealth generally lower for higher-income households; magnitudes are small (e.g., 0.003*** in core), consistent with stock MPCs being small since stocks are concentrated among higher-income households.

H3: Implications of demographic changes
- Demographic trends (PSID and U.S. Census projections) point to population aging, declining homeownership and changing educational and marital status shares.
- Estimated direct effect: population aging is estimated to lower the housing MPC by about 0.3 cents out of every additional dollar in net home wealth every decade.
- Overall implication: underlying demographic trends point to potentially lower aggregate MPCs out of wealth going forward.

### V. Robustness checks (appendix summary)
- Robustness across specifications:
  - Results for net housing wealth MPC (0.043***) robust to addition of household-specific controls, time fixed effects, household fixed effects, and composition adjustments (Table A1).
  - Stock holdings MPCs remain small and not significant across robustness checks (Table A2).
  - Total net wealth MPCs (0.003***) robust across alternative specifications (Table A3).
- Additional robustness:
  - Deflating using CPI rather than PCE leaves estimation results virtually unchanged.
  - Truncated samples excluding lowest and highest 1 percent of income distribution examined; interaction terms and group dummies used to probe non-linearities.

### VI. Quantitative implications and conclusion
- Representative quantitative results:
  - Net housing wealth MPC: 0.043***.
  - Gross home assets MPC: 0.062***.
  - Gross home-related debt MPC: 0.131***.
  - Total net wealth MPC: 0.003***.
  - Stock holdings MPC: 0.002 (not statistically significant in core estimates).
- Aggregate implication:
  - Using estimated MPCs across income groups, a sustained 10 percent fall in house prices would lead to a fall in aggregate consumption of about 1-1.4 percent.
  - Recent increases in financial asset values appear to have done little to support aggregate consumption because MPCs out of financial and stock wealth are small or insignificant.
- Policy-relevant inferences:
  - Housing wealth fluctuations have meaningful consumption effects, amplified by borrowing/collateral channels.
  - Heterogeneity by age, income, employment, and credit access matters for transmission of house-price shocks to consumption.
  - Demographic trends (aging) are likely to dampen the sensitivity of consumption to wealth going forward.

*Source: IMF staff calculations based on the Panel Study of Income Dynamics (PSID) and aggregate time series; estimation period 1999–2017 as reported in wpiea2019115.*

### REFERENCES

### REFERENCES

### Empirical studies on housing wealth and consumption
- Aladangady, A., 2017, “Housing wealth and consumption: evidence from geographically linked microdata”, American Economic Review, 107 (11), pp. 3415-3446.  
- Benjamin, J.D., Chinloy, P. and Jud, G.D., 2004, “Real estate versus financial wealth in consumption,” Journal of Real Estate Finance and Economics, 29 (3), pp. 341–354.  
- Bostic, R., and Surette, B., 2001, “Have the doors opened wider? Trends in family homeownership rates by race and income,” Journal of Real Estate Finance and Economics, 23, pp. 411–434.  
- Bostic, R., Gabriel, S., and Painter, G., 2009, “Housing wealth, financial wealth and consumption: new evidence from micro data,” Regional Science and Urban Economics, 39, pp. 79–89.  
- Bover, O., 2005, “Wealth effects on consumption: microeconometric estimates from the Spanish survey of household finances,” Documentos de Trabajo No. 0522, Banco de España.  
- Campbell, J., and Cocco, J., 2007, “How do house prices affect consumption? Evidence from micro data,” Journal or Monetary Economics, 54, pp. 591–621.  
- Carroll, C., Otzuka, M., and Slacalek, J., 2011, “How Large are Housing and Financial Wealth Effects? A New Approach,” Journal of Money, Credit and Banking, 43 (1), pp. 55–79.  
- Case, K.E., Quigley, J.M., and Shiller, R.J., 2005, “Comparing wealth effects: the stock market versus the housing market,” The B.E. Journal of Macroeconomics, 5 (1), pp. 1–34.  
- Cooper, D., 2013, “House price fluctuations: the role of housing wealth as borrowing collateral,” Review of Economics and Statistics, 95 (4), pp. 1183-1197.  
- Mian, A., Rao, K., and Sufi, A., 2013, “Household balance sheets, consumption, and the economic slump,” Quarterly Journal of Economics, 128, pp. 1687-1726.  
- Paiella, M, 2007, “Does wealth affect consumption? Evidence for Italy,” Journal of Macroeconomics, 29, pp. 189–205.  
- Paiella, M., 2009, “The stock market, housing and consumer spending: A survey of the evidence on wealth effects,” Journal of Economic Surveys, 23 (5), pp. 947–973.  
- Poterba, J., and Samwick, A., 1995, “Stock Ownership Patterns, Stock Market Fluctuations, and Consumption,” Brookings Papers on Economic Activity, 1995 (2), pp. 295–372.  
- Juster, T., Lupton, J., Smith, J., and Stafford, F., 2006, “The decline in household saving and the wealth effect,” Review of Economics and Statistics, 88, pp. 20–27.  
- Maki, D. and Palumbo, M., 2001, “Disentangling the wealth effect: a cohort analysis of household savings in the 1990s,” FEDS Working Paper No. 2001-23, Board of Governors of the Federal Reserve.  
- Parker, J, 1999, “Spendthrift in America? On two decades of decline in the US saving rate,” in B. Bernanke and J Rotemberg (eds), NBER Macroeconomics Annual. Cambridge, MA: MIT Press.  

### Theoretical foundations and consumption theory
- Attanasio, O. and Weber, G., 1994, “The UK consumption boom of the late 1980s: aggregate implications of microeconomic evidence,” Economic Journal, 104, pp. 1269–1302.  
- Campbell, J.Y. and Mankiw, N.G., 1990, “Permanent income, current income, and consumption,” Journal of Business and Economic Statistics, 8, pp. 269–279.  
- Carroll, C., and Kimball, M., 1996, “On the Concavity of the Consumption Function,” Econometrica, Vol. 64, No. 4, pp. 981–992.  
- Hall, R.E., 1978, "Stochastic Implications of the Life Cycle—Permanent Income Hypothesis: Theory and Evidence," Journal of Political Economy, 86, pp. 971–987  
- Modigliani, F. and Brumberg, R.H., 1954, “Utility Analysis and The Consumption Function: An Interpretation of Cross-section Data”, in Kenneth K. Kurihara, ed., Post‑Keynesian Economics, New Brunswick, NJ. Rutgers University Press, pp. 388–436.  
- Zeldes, S., 1989, “Optimal consumption with stochastic income: deviations from certainty equivalence,” Quarterly Journal of Economics, 104, pp. 275–298.  
- Keynes, J.M., 1936, “The General Theory of Employment, Interest and Money,” Palgrave Macmillan, Cambridge.  

### Methodology, econometrics, and surveys
- Engle, R.F. and Granger, C.W., 1987, “Co-Integration and Error Correction: Representation, Estimation, and Testing,” Econometrica, 55 (2), pp. 251–76.  

*Source: wpiea2019115 - REFERENCES*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019115.pdf_