Why Partial Insurance Fails Poor Households – New Model

A new working paper from arXiv lays out the math behind poverty traps and why standard insurance often makes them worse. The model shows that when a household loses a random chunk of its wealth, partial insurance coverage can actually deepen the trap instead of breaking it. The paper is titled "On the Impact of Insurance on Households Susceptible to Random Proportional Losses."

The core insight is simple. A household that loses 30% of its assets in a bad year needs a 43% gain just to get back to the starting point. If the loss is random and recurring, the household never builds the compounding base needed to cross the poverty threshold. The authors model this as a stochastic process where wealth follows a random walk with a reflecting barrier at the poverty line.

The Core Mechanism: Proportional Loss and the Poverty Threshold

The paper defines a household's wealth as a stock that grows through savings and shrinks through random proportional losses. When the loss is large enough to push wealth below a critical level, the household enters a "poverty trap" zone. In that zone, the expected growth rate is negative. The household has to divert income to basic consumption rather than productive investment.

The key variable is the loss fraction. A 10% loss might be survivable. A 40% loss can be catastrophic. The paper shows that the probability of being trapped rises sharply once the loss fraction exceeds the household's savings rate. For a household saving 15% of income, a 20% loss is manageable. A 30% loss is not.

The paper uses a calibrated simulation based on household survey data from low-income countries. The authors set the poverty threshold at $2 per day per capita, the savings rate at 10–20% of income, and the loss distribution as a lognormal with a mean of 15% and a standard deviation of 10%. The simulation runs 10,000 households over 50 years.

Why Partial Coverage Backfires

Insurance products typically cover a fixed percentage of the loss – 50% or 80%. The paper shows that this partial coverage can be worse than no coverage at all. The reasoning is mechanical. If the insurance payout is too small to restore wealth above the threshold, the household remains trapped. The insurance premium, meanwhile, reduces the household's savings capacity, making it harder to escape the trap through organic growth.

The authors model this as a trade-off. Full insurance covering 100% of the loss prevents the trap almost entirely. Partial insurance covering 50% reduces the frequency of trap entry a little. It increases the duration of each trap episode. The net effect for households with low savings rates is negative.