240 independent households each start with $10,000 in savings and $8,000/yr of surplus income. Every simulated year, each household may suffer 0, 1 or more insurable loss events (fire, theft, liability, etc.), and the simulator settles each claim exactly as a real policy would:
Claim count ~ Poisson(λ)
Severity S ~ Lognormal(μ, σ), mean ≈ $3,000
Insurer pays = clamp(S − Deductible, 0, Limit)
Out-of-pocket = S − Insurer pays
Wealth(t+1) = Wealth(t) + Income − Premium − Σ Out-of-pocket
The annual premium is not a slider — it is priced live from your own Deductible/Limit choice using the actuarial fair-premium formula, then marked up by the safety loading:
Premium = λ · E[Insurer pays | D, L] · (1 + loading)
Raising the deductible shifts small losses back onto the household but lowers the premium (less expected payout for the insurer to price in) — this is the core deductible/premium trade-off in real insurance planning. Raising the loading shows how insurer overhead and profit margin erode the deal. Turning Insurance OFF removes the premium entirely but leaves every loss fully out-of-pocket — watch the ruin rate (households whose wealth falls below $0) spike, which is exactly the risk-pooling and diversification argument for buying coverage instead of self-insuring.
- Each of the 240 bars is one independent household — the spread across bars is genuine cross-sectional risk, not noise.
- Bar height = current wealth; color fades from green (healthy) to red (ruined, wealth ≤ $0, frozen).