Spend or Invest? Analyzing MPC Heterogeneity Across Three Stimulus Programs

What drives within-person variation in the use of windfall income? We employ granular transaction data gathered by an account aggregator application to study how personal circumstances drive cross-sectional and within-person variation in the marginal propensity to consume (MPC) across the three pandemic stimulus programs. We apply a robust imputation estimator to recover the distribution of spending responses and show that liquidity constraints and debt play a crucial role in explaining variation in MPCs over time. In particular, we document the substitution between consumption spending and debt repayments. Higher indebtendness is associated with lower consumption responses due to a direct crowding-out effect. We further document a modest effect of fiscal transfers on retail investments but a strong effect on market participation with a large gender gap in the propensity to invest.