The distribution of wealth and the marginal propensity to consume

In a model calibrated to match micro- and macroeconomic evidence on household income dynamics, we show that a modest degree of heterogeneity in household preferences or beliefs is sufficient to match empirical measures of wealth inequality in the United States. The heterogeneity-augmented model's predictions are consistent with microeconomic evidence that suggests that the annual marginal propensity to consume (MPC) is much larger than the roughly 0.04 implied by commonly used macroeconomic models (even ones including some heterogeneity). The high MPC arises because many consumers hold little wealth despite having a strong precautionary motive. Our model also plausibly predicts that the aggregate MPC can differ greatly depending on how the shock is distributed across households (depending, e.g., on their wealth, or employment status).