Question to my insurance friends:
The ideal characteristics of an insurable risk according to Rejda and McNamara in Principles of Risk Management and Insurance are:
• There must be a large number of exposure units.
• The loss must be accidental and unintentional.
• The loss must be determinable and measurable.
• The loss should not be catastrophic.
• The chance of loss must be calculable.
• The premium must be economically feasible.
Why do private insurers insure against hurricane damage, but most do not insure against flood damage?
Flood is much harder to model. The industry CAT modelers will sell you models for each peril but the quality of the hurricane one is much better than flood because hurricane lends itself to modeling moreso than flood, which is very sensitive to small geographic differences.
On top of that and partly as a result we as society have decided to heavily subsidize insuring flood risk via the NFIP, so we’ve largely nationalized the market.
Flood is certainly insurable, especially nowadays, but we have a regulatory environment that largely crowds out private flood insurance.
3 Likes
hurricane’s make a lot of roof, siding, window damage.
floods ruin the home.
that’s my guess.
Private insurers had issues with being able to model either historically. The big boom in coastal property exposure in FL, the gulf, and the SE coast started happening at a time when the US was relatively lucky in terms of low land falling hurricane frequency. Then when hurricane activity started heating up again, insurers realized their historical data was woefully inadequate and embraced modeling. It’s not been perfect by any means, but has helped tremendously in terms of evaluating exposure.
Modelers didn’t have the data and computing power to evaluate flood exposure historically, but have made tremendous leaps in recent years in terms of remote sensing, hydrologic + hydraulic modeling, and computing power. Private insurers likely will move into this space more over time as they get their arms around modeled exposure, although consumers may not be willing to pay for the true cost of risk.
2 Likes
The chance of loss being calculable is a huge issue for flood, but I’d also throw in “what’s the maximum severity of the event giving rise to the loss.” Hurricanes have a theoretical maximum strength, but that’s not going to be found everywhere a hurricane strikes. There’s enough of a history to get an idea of the chance of a hurricane at a given strength, and we can model out how that changes with ocean temperatures.
Flood probability isn’t easily estimable even given history in an area. A certain location may have a history of flood events, but it’s got say severity S; a different location has only 1-2 flood events in its history, but one of them was 3S. Does the first location have the potential for a 3S flood event? If so, what’s the chance of it happening? What’s the chance of the second location having “just” an S flood event? Another 3S flood event? Did it have S events, but they were inconsequential so they’re not in the historical record?
Now throw in an event for Cities A, B and C that are 30 miles apart from each other and one of them may or may not be impacted further depending on whether one of the other two are part of the event. What’s the probability of those events? The probability of those events where the severities in one location do / don’t impact another? Then layer on where the water levels of rivers/lakes in the area are when the event occurs, and how that mitigates / aggravates flood damage.
Even if you can model a given flood event, assigning probabilities for certain situations is blindly throwing darts at a wall. That’s where you end up with the imbalance between insurers are comfortable with modeled losses such that they’ll write policies and resulting premiums are affordable for prospective insureds.
That might need to be rephrased as “…most in the United States do not insure…”
In this day and age, it is likely that American insurers would write flood insurance outside known zones of higher flood risk but-for the availability of the NFIP, and the perception that outside those known zones flood coverage isn’t needed.
And it is worth noting that flood is a thing in some commercial property programs, and even specialty homeowners programs…although being outside zones A* and V* are usually prerequisites for eligibility.
A lot of great points above. I just want to add another dimension to the issue. Insurers may provide insurance even when some of the requirements for an insurance risk to.be covered are not met if they think it will be profitable. Flood insurance in flood prone areas is clearly unprofitable. Hurricane insurance (and wild fire) was initially profitable until it was not but because it was provided before there is anchoring bias. Regulators and clients see it as standard rather than the aberration it is. Agriculture is another tricky cover if the insured waits for rainfall rather than irrigates in places were rainfall is unpredictable.