Flood Insurance

Study guides: Exam 6C, Exam 6U

Flood Insurance covers damage from overland water — river, coastal and surface flooding — as distinct from sewer backup or water escaping from plumbing. Until roughly 2015 residential overland flood was uninsurable in Canada: it was excluded from every homeowner policy, and losses were met by ad hoc government payments under DFAA. It is now widely offered, but not to the highest-risk properties, which is the coverage gap Canadian flood policy is trying to close.

• Why it was uninsurable. Flood violates the conditions for private insurability: losses are spatially correlated (one event hits every property in the floodplain simultaneously), severe Adverse Selection means only floodplain residents would buy, and the risk was un-mappable at the property level without modern elevation and hydrological data.
• What changed: high-resolution flood mapping and catastrophe modelling made property-level pricing possible; the 2013 Alberta and Toronto floods made the coverage gap politically visible; and reinsurers made capacity available. Insurers now offer overland flood as an endorsement, usually with a sub-limit, a separate deductible and eligibility screening.
• The residual gap. Roughly the highest-risk properties — those in frequent-flood zones — remain effectively uninsurable at any premium a homeowner would pay. Proposals for a national flood insurance program centre on a public high-risk pool with a federal backstop, private delivery, and risk-based pricing with subsidy made explicit rather than hidden.
• Moral Hazard at the land-use level is the defining policy problem: government disaster assistance after every flood, and any subsidised premium, both reduce the incentive to avoid building in floodplains or to invest in mitigation. A well-designed program prices the risk and subsidises the household, not the location.
• Actuarial features: loss distributions are extremely heavy-tailed and correlated, so pricing depends on catastrophe models rather than experience; the PML drives reinsurance and capital, and OSFI treats flood accumulation as a concentration and climate issue.
• Climate change is shifting the hazard itself, so the historical record understates the future — a non-stationarity problem that makes conventional return-period language unreliable.

Why Voluntary Flood Coverage Cannot Be Priced at an Average A community of 10{,}000 homes has three zones: 500 homes with an annual flood probability of 10\% and average loss \$80{,}000; 1{,}500 homes at 1\% and \$60{,}000; 8{,}000 homes at 0.05\% and \$40{,}000.

Compute the risk-based premiums and the community-average premium, and explain what happens if the average is charged.

Answer Risk-based pure premiums:

\begin{align*}
\text{Zone 1} &= 0.10 \times \$80{,}000 = \$8{,}000 \\
\text{Zone 2} &= 0.01 \times \$60{,}000 = \$600 \\
\text{Zone 3} &= 0.0005 \times \$40{,}000 = \$20
\end{align*}

Community average, weighting by home count:

\begin{align*}
\bar{P} &= \frac{500(\$8{,}000) + 1{,}500(\$600) + 8{,}000(\$20)}{10{,}000} \\
&= \frac{\$4{,}000{,}000 + \$900{,}000 + \$160{,}000}{10{,}000} \\
&= \$506
\end{align*}

Charge \$506 to everyone and the pool collapses. Zone 3 homeowners face a premium 25 times their expected loss and do not buy. Zone 2 homeowners are near break-even and mostly do not buy. Zone 1 homeowners face a premium one sixteenth of their expected loss and all buy — and the pool now needs \$8{,}000 per policy against \$506 collected. This is Adverse Selection running to completion in a single step.

What actually works:

• Price to the zone. Zone 3 buys at \$20 (trivially cheap, so take-up is high), Zone 2 buys at \$600. Both are insurable privately.
• Zone 1 is the real problem. At \$8{,}000 nobody buys; below \$8{,}000 no insurer writes. This is where a public high-risk pool with an explicit, funded subsidy and a mitigation or relocation requirement is the only structure that works — and it should be paired with land-use rules preventing the zone from growing.