Estimating the Marginal Propensity to Consume Using the Distributions of Income, Consumption and Wealth
Estimating the Marginal Propensity to Consume Using the Distributions of Income, Consumption and Wealth (DOI: https://doi.org/10.29412/res.wp.2019.04) was written by Jonathan Fisher, David Johnson, Timothy Smeeding, and Jeffrey Thompson in 2019. It was published as a part of the Federal Reserve Bank of Boston Research Department Working Papers series (no. 19-4).
The authors estimate the marginal propensity to consume (MPC) using the PSID.
- PSID measures of income and consumption exclude government- and employer-provided health benefits.
- Authors adjust values to 2013 dollars and scale to square root of family size.
- Taxable, transfer, and Social Security income is summed across all household members. Post-tax value is estimated using the Kimberlin, Kim, and Shaefer (2015) model.
- All categories of wealth are summed across all household members.
- Only categories of consumption that were available pre-2003 (food, housing, transportation, education, and child care) are summed across all household members.
- Use longitudinal data 1999-2013: 35,000 cases
Haig-Simons equation:
Y = C + ∆W
where Y is income, C is consumption, and W is wealth.
The authors estimate the Euler equation:
∆Ci,t-2 = α + β ∆Yi,t-2 + δ Zi,t-2 + ρ1 statei,t-2 + ρ2 yeari,t-2 + εi
where the ∆ terms are computed as log difference between years t and t-2. As the equation indicates, state and year fixed effects are controlled for. The longitudinal dataset is used with longitudinal weights, and repeat families have clustered errors.
The authors then repeat the model with an interaction between income and wealth quintile. Quintiles are computer annually, specifically referencing year t-2.
The authors find a lower MPC than prior work, in part due to the fact that they measure change over a 2 year interval (longer than conventional). They also find that MPC is lower for higher wealth households, suggesting that they are more capable of smoothing income.
The authors connect this to Krueger (2012), which found that an income transfer from high-wealth households to low-wealth households would drive consumption higher.