Household Heterogeneity ABM Model
How do heterogeneous household consumption rules, saving buffers, and wealth positions shape the aggregate demand response to fiscal, monetary, and distributional shocks - and why does the shape of the income/wealth distribution matter more than the mean?
Graph workspace
Multi-panel time-series with wealth distribution sidebar
Household heterogeneity ABMs produce stochastic time series (aggregate output, consumption, unemployment) alongside cross-sectional distributions (wealth distribution, MPC distribution) that shift endogenously over time. The canonical visualization pairs a time-series panel showing aggregate demand dynamics with a histogram of the wealth distribution and a bar chart of MPC by wealth quintile. This makes the connection between distributional state and aggregate response visible in one view.
Current results
Output Index
Aggregate output relative to period-0 baseline of 100. Rises when aggregate consumption increases, falls during recessions driven by unemployment or demand shortfalls.
85.88
index
Unemployment Rate
Fraction of households not employed. Rises when firms cut labor in response to weak demand; falls when aggregate consumption drives hiring.
27.00%
percent
Wealth Gini
Gini coefficient of the household wealth distribution at period . Evolves endogenously as income shocks, consumption, and saving reshape the cross-section.
0.1295
coefficient (0-1)
Aggregate MPC
Economy-wide marginal propensity to consume. Higher when more households are in the constrained (low-wealth) regime. The key summary statistic for fiscal multiplier analysis.
0.7080
ratio
Controls
Share of extra income spent by households already at their savings target; the model's lowest rate.
A higher raises the consumption share of output and the Aggregate MPC readout, lifts the whole MPC-by-quintile bar chart, and leaves average unemployment lower at a high than at a low , though not at every step (it rises from 0.65 to the 0.70 default, and from 0.85 to 0.90). It does not raise the output index in this sketch: also sets how strongly the simulated cycle feeds back on itself, so across the slider output swings rather than climbs, and its run average is lower at the 0.7 default than at . It also compresses the achievable MPC gradient, because the ratio between the most and least constrained household is capped at - 1.43x at and the default , but 7x at with at its 0.6 maximum.
How much being short of the savings target raises a household's spending rate; zero means no difference.
A higher widens the MPC range the rule can span, from at to , so whatever gradient exists between low-wealth and high-wealth quintiles is steeper. The parameter is also what makes the fiscal multiplier depend on the wealth distribution at all: at the multiplier is distribution-invariant. Whether a gradient is visible in this sketch depends on households actually sitting at different distances from their targets, which the seed does not deliver.
How concentrated wealth is at the start; it barely changes how much of their income households spend.
Moves the starting wealth histogram. The Wealth Gini readout is an end-of-run value, and by then wealth has converged toward each household's target, so it barely changes with this slider. It has no reliable path to the MPC-by-quintile chart or to the size of the transfer response: re-concentrating wealth does not lift households across their own buffer-stock targets in this sketch, and by the transfer period the wealth process has in any case pulled households back toward regardless of where they started. Do not read a multiplier difference off this slider.
Transfer amount: 0
Average one-time cash payment per household, paid once in period 20; set to zero for none.
Positive transfer creates a one-quarter output spike at period 20 - its height is the impact MPC times the transfer as a share of baseline output - which then decays as recipients rebuild buffers at speed . What targeting the transfer at low-wealth households can add is bounded above by whatever the seed; whether this sketch actually shows any targeting gain is a separate matter, and it does not, because the wealth distribution has converged toward by the transfer period.
Unemployment benefit rate: 0.4
The share of a worker's wage paid to them while they are unemployed.
Higher benefit rate stabilizes output by preventing consumption collapse among the unemployed. Reduces the amplitude of business cycle fluctuations in the output panel.