Working Paper: NBER ID: w18883
Authors: Marianne Bertrand; Adair Morse
Abstract: Have rising income and consumption at the top of income distribution since the early 1980s induced households in the lower tiers of the distribution to consume a larger share of their income? Using state-year variation in income level and consumption in the top first quintile or decile of the income distribution, we find evidence for such "trickle-down consumption." The magnitude of effect suggests that middle income households would have saved between 2.6 and 3.2 percent more by the mid-2000s had incomes at the top grown at the same rate as median income. Additional tests argue against permanent income, upwardly-biased expectations of future income, home equity effects and upward price pressures as the sole explanations for this finding. Instead, we show that middle income households' consumption of more income elastic and more visible goods and services appear particularly responsive to top income levels, consistent with supply-driven demand and status-driven explanations for our primary finding. Non-rich households exposed to higher top income levels self-report more financial duress; moreover, higher top income levels are predictive of more personal bankruptcy filings. Finally, focusing on housing credit legislation, we suggest that the political process may have internalized and facilitated such trickle-down.
Keywords: Income Inequality; Consumption; Household Savings; Credit Supply; Behavioral Economics
JEL Codes: D1; E21
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Income levels at the top are exogenous to the consumption decisions of lower-tier households (D10) | Consumption of non-rich households (D12) |
| Consumption patterns of the wealthy (E21) | Consumption response of non-rich households (D12) |
| Local price pressures and social comparisons (P22) | Consumption response of non-rich households (D12) |
| Rising top income levels (D31) | Future income for non-rich households (G59) |
| Increase in income levels among the top quintile or decile of the income distribution (D31) | Increase in consumption among middle-income households (D12) |
| 10 percent rise in the income of the top 90th or 80th percentile within a state (D31) | Increase in consumption among middle-income households (D12) |